Objective: To determine whether higher-intensity treadmill walking during rehabilitation in those newly able to walk after stroke is feasible, is detrimental to walking or is beneficial. Design: A single-blind, randomized trial. Setting: Two rehabilitation units. Participants: Thirty people with first stroke. Interventions: Experimental group received 30 minutes of higher-intensity treadmill walking, three times per week for six weeks, in addition to usual physiotherapy. Control group received usual physiotherapy only. Main outcome measures: Feasibility was measured by examining compliance and adverse events. Detriment to walking was measured by examining pattern and quality. Benefit to walking was measured as capacity using six-minute walk test and speed, at baseline (Week 0), immediately after (Week 6) and at three months (Week 18). Results: Overall compliance was 89%, with no adverse events reported. There was no between-group difference in walking pattern and quality. By Week 6, the experimental group improved walking capacity by 62 m (95% confidence interval (CI) 10 to 114), comfortable walking speed by 0.18 m/s (95% CI 0.07 to 0.29) and fast walking speed by 0.18 m/s (95% CI 0.03 to 0.35) more than the control group. By Week 18, the experimental group was still walking 0.26 m/s (95% CI 0.12 to 0.41) faster than the control group. Conclusions: Higher-intensity treadmill walking during rehabilitation after stroke is feasible and not detrimental to walking pattern and quality in those newly able to walk. 94 participants are required to detect a between-group difference of 75 m on six-minute walk test (a 0.05, b 0.90) in future trials.