Background: People with dementia living in institutionalised residential aged care facilities (RACFs) may manifest behavioural and psychological symptoms of dementia (BPSD), such as agitation and apathy, resulting from their unmet needs. These unmet needs are often due to inappropriate environmental conditions, including excessive or insufficient levels of stimulation in the environment and limited opportunities for engagement. In response to these shortcomings, there has been a recent trend to improve the quality of RACF environments through the de-institutionalisation of facilities by the creation of dementia-friendly environments (DFEs) incorporating outdoor natural landscapes, such as gardens, into RACFs. Garden elements, for example, green spaces and blue spaces, have been shown to have positive impacts on the health and well-being of individuals, and this includes for older people with dementia. In the past few years, several new initiatives have been attempted to enable people with dementia to take advantage of the potential benefits of gardens by encouraging residents to spend more time in the gardens and to actively engage with garden-related activities. It has been found that to maximise the potential benefits of gardens for people with dementia, eight DFE characteristics can be incorporated into their design. However, to date, there is very little evidence about the positive outcomes of a garden that has been created with DFE characteristics on agitation, apathy and engagement for older people with dementia living in RACFs. This is due to both the paucity of studies and the limitations of previous studies, for example, the use of weak methodological approaches as well as the lack of environmental assessment of the gardens and the exclusion of an architectural design phase. Therefore, this current study aims to fill the gaps in the existing studies. Objectives: This research aims to determine the effect of a garden which had been improved according to the DFE characteristics on the level of agitation, apathy, and engagement of people with dementia living in a RACF in Queensland. Methods: This study utilised a single case study design involving two units of analysis: (1) a garden and (2) 10 people with dementia living in one RACF. This study is organised into three phases. Phase One consisted of an environmental assessment of the selected RACF garden via the Dementia Therapeutic Garden Audit Tool and an interview with the gardener of the facility, as well as the recruitment of 10 participants and the compiling of their social biographies in order to understand the participants’ needs and interests. In Phase Two, the chosen garden, Ruby Garden, improvements were identified which addressed the DFE deficiencies found during the environmental assessment in Phase One. Following this, a 3D model of the garden was created to incorporate and visualise the DFE characteristics. After the garden beds were prepared, the items consisting of softscapes and hardscapes were incorporated into the Ruby Garden based on the model created. In Phase Three, the participants were taken into the garden to experience the improved garden for 20 sessions, 5 days a week for 4 weeks, each session lasting 60 minutes. During this period, the participants’ levels of agitation, apathy and engagement was assessed. Assessment of agitation was conducted pre- (Week 0) and post-intervention (Week 5). On-site observations of engagement and apathy were conducted pre-intervention (Week 0) over one session and once a week during the 4-week intervention period except for Week 1, when two assessments took place (i.e., beginning (Week 1-1) and end (Week 1-2) of the week). At Week 8, a follow-up assessment was undertaken to determine whether the intervention effects on agitation, engagement and apathy were sustained over time. Participants’ activities in the garden were also observed every day at 10-minute intervals over the 4-week intervention period from the time they entered the garden until they left. Post-intervention qualitative interviews with the 10 participants with dementia and 10 staff were also conducted to seek their views on the improved garden. Results: The results of the environmental assessment via the Dementia Therapeutic Garden Audit Tool carried out in Phase One showed the original garden to be a poor environment for people with dementia with more than half of the assessing questions (39 out of 74 questions) scoring 0-1 (i.e., poor). The Ruby Garden was found to be successful (score of 3) in only 21 out of 74 questions. The lack of the eight DFE characteristics, particularly those concerning sensory stimulation and meaningful engagement, was noted in the Ruby Garden original design. Participants also highlighted the need for these two characteristics while their social biographies were being collected. In Phase Two the results of the post-design assessment of the garden showed that the garden was improved from the perspectives of all eight DFE characteristics. In total, the evaluation showed an increase in the total number of questions scoring very good (score of 3), from 21 for the environmental assessment phase (i.e., Phase One), to 36 for the post design assessment of the garden (i.e., Phase Two). Likewise, sensory stimulation and meaningful engagement, which were identified as the characteristics with the greater number of poor scores in Phase One, were improved in the number of questions scoring 2 or 3 (i.e., moderately good or very good) from 3 to 12 (out of 12) and from 0 to 4 (out of 7), respectively. In Phase Three the quantitative findings of the assessment of agitation of participants at Week 0 (baseline), Week 5 (post-intervention) and Week 8 (follow-up), assessed through the Cohen Mansfield agitation inventory-short form (CMAI-SF), demonstrated no significant improvement in the level of agitation following visits to the improved garden (χ 2 (2) = 5.167, p = .076). However, the results showed an increase in the level of participants’ engagement and a reduction in their level of apathy following the intervention. The statistical analysis showed a significant difference in participants’ level of engagement between Week 0 (before intervention) and during the intervention (Week 1-1, p < .01; Week 1-2, p < .01; Week 2, p < .01; Week 3, p < .05; Week 4 p < .05), where the level of engagement was higher than that before the intervention. Furthermore, the assessment of participants’ level of engagement at follow-up (Week 8) demonstrated a significantly lower level of engagement when compared to the 4 weeks of intervention (Week 1-1, p < .01; Week 1-2, p < .05; Week 2, p < .05; Week 3, p < .05; Week 4 p <.05). Similarly, the statistical analysis found a higher level of apathy in participants at Week 0 (before intervention) when compared to during the intervention (Week 1-1, p < .05; Week 1-2, p < .01; Week 2, p < .05; Week 3, p < .01; Week 4 p < .01). A significant difference was also found in participants’ level of apathy when comparing Week 8 (follow-up) and during Week 1-2 (p < .05), Week 2 (p < .05), Week 3 (p < .01) and Week 4 of the intervention (p < .05), where the level of apathy was higher at follow-up than during the intervention. A combination of deductive (concept-driven) and inductive (data-driven) thematic analysis was employed to analyse the qualitative data. Three themes were developed via the deductive approach: the presence of sensory provoking elements in the garden, meaningful engagement in the garden, and accessibility of the garden, which indicated that the improved garden was successful in providing such DFE characteristics. Moreover, two themes emerged from the inductive approach: garden impacts and garden experiences. The findings of the inductive approach showed that the improved garden had several benefits for people with dementia, staff members and families. The garden was perceived by the interviewees to improve the mood and restless behaviours of people with dementia as well as the mood of the staff members and families. The inductive qualitative results also indicated that the majority of people with dementia and other residents had a good perception and positive feedback on the garden. Nevertheless, they recommended some solutions for improving the garden, which need to be considered in the future development of the garden. Conclusions: The results of this case study demonstrated that the garden, which was improved in line with DFE characteristics, had positive outcomes for people with dementia. This finding validates and supports several previous studies reviewed in the area of creating such gardens, which showed the positive outcomes of the gardens for people with dementia. This study is the first known case study that comprehensively assessed the environment before the design, collected the participants’ social biographies and also created a 3D model for improvement of the garden. Also, it is the first intervention study that evaluated the impacts of an enhanced garden on the level of apathy as an outcome measure and it also showed promising results. Thus, this study can lay a foundation for future studies in this area.