Background: It has been reported that following rehabilitation, only 7% of stroke survivors are able to walk at a level commensurate with community participation. Previous research indicates that treadmill and overground walking training can improve walking capacity in people living in the community after stroke. The main objectives of the AMBULATE trial are to determine (i) whether a 4-month treadmill walking program is more effective than a 2-month program, compared to control, in improving walking capacity, health and community participation and (ii) the "threshold" walking speed that results in sufficient walking capacity that makes walking self-sustaining. Methods/Design: A prospective randomised controlled trial of unsupported treadmill training with a 12 month follow-up with concealed allocation and blinded assessment will be conducted. 210 community-dwelling people after stroke who are able to walk independently but slowly will be recruited and randomly allocated to either a 4 month training group, 2 month training group or the control (no intervention) group. Intervention for the two training groups will occur 3 days per week for 30 minutes each session. Measurements of walking, health and community participation will be taken at baseline, 2 months, 4 months, 6 months and 12 months. This study has obtained ethical approval from the relevant Human Research Ethics Committees. Discussion: By improving stroke survivors' walking ability, it is likely also to improve their general wellbeing by promoting better health and greater community participation. Furthermore, if stroke survivors can reach a point where their walking and community participation is self-sustaining, this will reduce the burden of care on family and friends as well as the economic burden on the health system. Given the major demographic shift in developed nations involving significant growth in the aged population, this research will make an important evidence-based contribution to the promotion of healthy ageing.