Abstract Mental health problems are associated with lower quality of life, increased unscheduled care, high economic and social cost, and increased mortality. Nature-based interventions (NBIs) that support people to engage with nature in a structured way are asset-based solutions to improve mental health for community based adults. However, it is unclear which NBIs are most effective, or what format and dose is most efficacious. We systematically reviewed the controlled and uncontrolled evidence for outdoor NBIs. The protocol was registered at PROSPERO (CRD42020163103). Studies that included adults (aged ≥18 years) in community-based settings with or without mental and/or physical health problems were eligible for inclusion. Eligible interventions were structured outdoor activities in green and/or blue space for health and wellbeing. We searched ASSIA, CENTRAL, Embase, Greenfile, MEDLINE, PsycINFO, and Web of Science in October 2019; the search was updated in September 2020. We screened 14,321 records and included 50 studies. Sixteen studies were randomised controlled trials (RCTs); 18 were controlled studies; and 16 were uncontrolled before and after studies. Risk of bias for RCTs was low to moderate; and moderate to high for controlled and uncontrolled studies. Random effects meta-analysis of RCTs showed that NBIs were effective for improving depressive mood −0.64 (95% CI: 1.05 to −0.23), reducing anxiety −0.94 (95% CI: 0.94 to −0.01), improving positive affect 0.95 (95% CI: 0.59 to 1.31), and reducing negative affect −0.52 (95% CI: 0.77 to −0.26). Results from controlled and uncontrolled studies largely reflected findings from RCTs. There was less evidence that NBIs improved physical health. The most effective interventions were offered for between 8 and 12 weeks, and the optimal dose ranged from 20 to 90 min. NBIs, specifically gardening, green exercise and nature-based therapy, are effective for improving mental health outcomes in adults, including those with pre-existing mental health problems.