Aim Rising mortality and disease prevalence in the homeless have been largely attributed to addiction disorders. This review aimed to assess whether Motivational Interviewing (MI) is effective in changing substance misuse behaviours in the homeless, specifically: 1. reducing substance misuse; 2. increasing addiction treatment linkage; and 3. whether MI effectiveness varied according to the different levels of social disadvantage within homeless populations. Method Electronic databases and other sources were searched (to July 2021) for relevant randomized trials and comparative studies. Risk of bias in included studies was evaluated using the Cochrane Risk of Bias tool. A Narrative Synthesis framework was applied to included studies. Moderator variables subgroup analyses were planned a priori. PROSPERO study protocol registration: CRD42019134312 Results The searches found 1885 records; after application of inclusion criteria n~=~11 studies from 30 articles were included in the review, all from the United States. There was a paucity of research regarding MI effectiveness for substance misuse outcomes in homeless populations, with a focus on short-term rather than long-term impacts. Risk of bias was generally low but was high for detection bias in most studies. MI appeared to be more effective overall amongst adult homeless persons, yielding consistently small effects, and alcohol use behaviours seemed to be more amenable to change as a result of MI/MET (Motivational Enhancement Therapy) interventions than drug use ones. Limited evidence with high risk of bias indicated that social gradient may attenuate MI effectiveness within the young homeless population, with no impact in the most disadvantaged. Conclusions The review’s mixed findings discourage the use of MI as a stand-alone substance use intervention in homeless populations. Although the review findings did not identify MI effectiveness for substance use according to the external level of social disadvantage faced by homeless persons, this should be a focus for further research.