The majority of existing studies of recreational preferences and forest characteristics focused on single stand attributes and demonstrated that people prefer stands with visual variation. However, it may be too simple since most people experience more than one stand when visiting a forest. This study aims at evaluating the effects of variation both within and between stands on recreational values. A choice experiment (CE) was applied to elicit people's preferences for forest types on their next recreational visit. Each alternative is presented with drawings of three forest stands which differ with respect to tree species, height (age) and distance to the site, the latter representing the cost factor – willingness-to-travel. Respondents also compose their ideal recreational forest by selecting three types of stands from the catalogue of drawings. We find that mixed tree species are preferred compared to monocultures. Stands with trees of varying height (uneven-aged stands) are preferred over stands consisting of trees of the same height (even-aged ones). Variation between stands is found to contribute positively to recreational value, and in some instances, this may outweigh contribution of variation within a stand. Comparing respondents' composition of their ideal forest with elicited preferences from the CE, confirm these findings.