Objective: The objective of the study is to evaluate the effectiveness of interventions to increase physical activity (PA) in 0-5 year olds and to determine what works, for whom, in what circumstances. Design: Systematic review, meta-analysis and realist synthesis. Data sources: Embase and EBSCOhost (Academic Search Complete, CINAHL Complete, Global Health, MEDLINE Complete, PsycINFO, SPORTDiscus with full text), up to and including April 2017. Eligibility criteria: Published in a peer-reviewed English language journal; randomized or controlled trial design; aimed to increase children's PA levels; reported on objectively assessed PA in children between 0 and 5.9 years at baseline and post-intervention. Results: Thirty-four studies were included in the review, mostly conducted in the preschool/childcare setting. Meta-analyses showed an overall non-significant (Z = 0.04, p = 0.97) mean difference of 0.03 (95% CI = -1.57, 1.63) minutes/day for light-intensity PA (n = 11). The overall mean difference for moderate-intensity to vigorous-intensity PA (n = 21) was 2.88 (95% CI = 1.54, 4.23) minutes/day, indicating a small but significant overall positive effect (Z = 4.20, p < 0.001). The realist synthesis provided insights into the key contexts and mechanisms that appeared to be effective at changing children's PA. Conclusion: Based on a quantitative and qualitative examination of the evidence, this review provides specific recommendations for effective early childhood PA interventions for practitioners and policymakers.