Questions: Is treadmill walking with body weight support during inpatient rehabilitation detrimental to walking quality compared with assisted overground walking? Does it result in better walking capacity, perception of walking or community participation? Design: Analysis of secondary outcomes of a randomised trial with concealed allocation, assessor blinding and intention-to-treat analysis. Participants: 126 patients unable to walk within 4 weeks of a stroke who were undergoing inpatient rehabilitation. Intervention: The experimental group undertook up to 30 minutes of treadmill walking with body weight support via an overhead harness per day while the control group undertook up to 30 minutes of overground walking. Outcome measures: The secondary outcomes were walking quality and capacity, walking perception, community participation and falls. Results: Six months after entering the study, there was no difference between the groups of independent walkers in terms of speed (MD 0.10 m/s, 95% CI -0.06 to 0.26) or stride (MD 6 cm, 95% CI -7 to 19). The independent walkers in the experimental group walked 57 m further (95% CI 1 to 113) in the 6 min walk than those in the control group. The experimental group (walkers and non-walkers) rated their walking 1 point out of 10 (95% CI 0.1 to 1.9) higher than the control group. There was no difference between the groups in community participation or number of falls. Conclusion: Treadmill training with body weight support results in better walking capacity and perception of walking compared to overground walking without deleterious effects on walking quality. Trial registration: NCT00167531.