The effectiveness of peer-mediated intervention (PMI) in teaching social skills to people with Autism Spectrum Disorder (ASD) has been attributed to a variety of factors, and the systematic literature reviews and meta-analyses that evaluated the effectiveness of PMI in teaching social skills to people with ASD restricted their analysis mainly to studies conducted with children or in inclusive settings. Considering this, this systematic review sought to update and expand knowledge about the relevant variables for the effectiveness of PMI in teaching social skills to people with ASD. A comprehensive search was conducted across the Scopus, Web of Science, and PubMed databases using the term "autism*" in combination with the terms "peer mediated*", "peer interaction*", "peer training*", and "peer mentoring". This search strategy yielded 70 relevant articles for further analysis. In each article, various aspects were analyzed, including but not limited to: characteristics of participants and peers, experimental environment, PMI procedures implemented, integrity measures of teaching procedures, and acquisition, generalization, and maintenance results. It was found that most studies: 1) were conducted in schools, with children with typically developing peers; 2) implemented the PMI using teaching packages consisting of Instructions, modeling, prompt, role-play and feedback; 3) showed effective or partially effective acquisition results, especially when measures of the integrity of the procedure implementation were performed. It was suggested the need for further studies to evaluate the effect of PMI on generalization and maintenance of social behaviors and its effect to establish or increase these behaviors in adolescents and adults with ASD. Additionally, some limitations to be addressed in future reviews on this topic are discussed.