Background Positive Youth Development (PYD) delivered outside school aims to enable young people to develop positive assets such as relationships and confidence, rather than to merely address risk. Existing reviews of PYD effects on substance use or violence are old and unsystematic. Objectives To systematically review evidence to answer the following questions: what theories of change inform PYD interventions addressing substance use and violence? What characteristics of participants and contexts are identified as barriers to and facilitators of implementation and receipt in process evaluations of PYD? What is the effectiveness and cost-effectiveness of PYD in reducing substance use and violence? What characteristics of participants and contexts appear to moderate, or are necessary and sufficient for, PYD effectiveness? Data sources A total of 21 bibliographic databases; 35 websites and contacting authors. Review methods We included reports published in English since 1985 and reporting on theories of change, as well as process, outcome and economic evaluations of PYD targeting 11- to 18-year-olds and addressing substance use or violence. References were screened on title/abstract and, where appropriate, on full report. Data extraction and quality assessment used Critical Appraisal Skills Programme, Evidence for Policy and Practice Information and Co-ordinating Centre and Cochrane tools. Theories of change and process evaluations were qualitatively metasynthesised. Outcome evaluations were synthesised narratively and meta-analytically. Results 32,394 unique references were identified and 48 were included. A total of 16 reports described theories, 13 (10 studies) evaluated processes and 25 (10 studies) evaluated outcomes. Theories of change PYD interventions aim to offer opportunities for young people to develop positive ‘assets’ such as skills and confidence. These are theorised to promote and be promoted by young people’s ‘intentional self-regulation’, which involves reflecting on behaviour; determining goals; using existing resources to pursue these; and redirecting effort when thwarted. This enables ‘developmental regulation’, namely individuals capitalising on other opportunities to promote personal development. Positive assets thus accrued reduce health risks by reducing the impact on individuals of environmental risk or by ameliorating the impact of such risks. The literature offers limited insights beyond these general ideas. Process evaluations Community engagement ensured that programmes were accessible and appealing. Staff capacity and continuity were crucial factors but often challenging when programmes could not offer full-time jobs. Tensions arose between a desire to empower participants to choose activities and a requirement for them to undertake a breadth of activities. Outcome evaluations Meta-analyses of all combined outcomes and of short-term alcohol use, illicit drug use and smoking found no significant effects. There were small, statistically significant, short-term effects for an omnibus measure of substance use and for violence. We could not undertake metaregression to assess sociodemographic moderators but narrative synthesis suggested no clear pattern of effects by sex. We found no economic evaluations. Limitations Insufficient studies precluded qualitative comparative analyses. Conclusions How PYD might promote health is currently undertheorised. Implementation can be challenging. We found little evidence that current PYD interventions delivered outside school reduce substance use or violence. However, these may not constitute a test of the effectiveness of the PYD model, as some included interventions that, although meeting our inclusion criteria, were not exemplars of PYD. Future work Further evaluations should assess interventions employing PYD theory of change. Study registration This study is registered as PROSPERO CRD42013005439. Funding The National Institute for Health Research Public Health Research programme.