Stroke rehabilitation for people living in the community is commonly delivered either in a centre, outpatient or day hospital setting. More recently, services may be offered in the actual home of the person as home-based or domiciliary rehabilitation. There are differing reports of the benefits and barriers of home-based vs. centre-based community rehabilitation. This systematic review sought to pool data from all retrieved studies that compared the functional benefits of home-based vs. centre for community-dwelling people with stroke. A comprehensive search strategy was implemented in all major databases (Cochrane library, Medline, AMED, Embase, Ageline, Cinahl, PEDro) for randomised controlled trials investigating this question in relation to functional benefits as a primary outcome and carer, cost or other benefits as secondary outcomes. There were no language or date limits. Eleven trials were found and results pooled for the Barthel Index, the measure of functional independence used consistently across the majority of retrieved studies. There was a significant effect in favour of home-based rehabilitation at 6 weeks (P=0.03) and 3-6 months (P=0.01). The effects were less clear at 6 months, although this was using the less sensitive version of the Barthel Index (P=0.27 or adjusted P=0.04). Individual studies reported cost benefits and increased carer satisfaction in favour of home-based rehabilitation. The provision of rehabilitation for people living in the community should trend towards home-based. Further research is required into adverse events and the experiences of all stakeholders.