Objective: This study aimed to determine effectiveness of current interventions to improve real world walking for people with stroke and specifically whether benefits are sustained. Data sources: EBSCO Megafile, AMED, Cochrane, Scopus, PEDRO, OTSeeker and Psychbite databases were searched to identify relevant studies. Review methods: Proximity searching with key words such as ambulat*, walk*, gait, mobility*, activit* was used. Randomised controlled trials that used measures of real world walking were included. Two reviewers independently assessed methodological quality using the Cochrane Risk of Bias Tool and extracted the data. Results: Nine studies fitting the inclusion criteria were identified, most of high quality. A positive effect overall was found indicating a small effect of interventions on real world walking (SMD 0.29 [0.17, 0.41]). Five studies provided follow-up data at > 3-6 months which demonstrated sustained benefits (SMD 0.32 [0.16, 0.48]). Sub group analysis revealed studies using exercise alone were not effective (SMD 0.19 [-0.11, 0.49]) but those incorporating behavioural change techniques (SMD 0.27 [0.12, 0.41]) were. Conclusions: A small but significant effect was found for current interventions and benefits can be sustained. Interventions which include behaviour change techniques appear more effective at improving real world walking habits than exercise alone.