This review investigated whether randomised controlled trials attempting to improve sleep or prevent sleep problems in 0–5 year olds influenced nocturnal sleep duration, day-time naps, or 24-h sleep. Medline (Ovid), EMBASE, and CINAHL were searched from inception until 9 July 2020 and supplemented with hand searching. Search results were screened, eligible data were extracted, and risk of bias was assessed by at least two reviewers. Of 8571 publications considered, 32 trials which used a variety of subjective and objective sleep measurements were included in generic inverse variance random effects meta-analysis of nocturnal (n = 24), day-time (n = 14), and 24-h (n = 13) sleep duration. Overall, sleep interventions increased nocturnal sleep duration by a mean of 9 min (95% CI 4.1 to 13.8, I²28%) per night when compared with no sleep intervention. Increases were predominantly seen in sleep-only, rather than multi-component interventions. Total 24-h sleep duration tended to increase by a similar amount (8.6 min (95% CI -2.7 to 19.8, I² = 59%)), but this was mainly only seen in studies that assessed sleep using diaries. There was no evidence that interventions changed day-time sleep duration. Future studies should involve sleep-only rather than multi-component interventions, and use objective sleep measures (reviewregistry857).