INTRODUCTION: Up to 80% of children and adolescents with autism spectrum disorders (ASD) has a sleep disorder, in whose genesis some alterations in the regulation of melatonin are apparently involved. The aim of this meta-analysis was to determine the efficacy and safety of melatonin for management of certain sleep disorders in children with ASD. METHODS: Three reviewers extracted relevant data from high-quality double-blind randomized clinical trials published in primary, clinical trials, systematic reviews and gray literature databases; further search was conducted in snowball. Data were analyzed with RevMan 5.3. An inverse of variance analysis by random-effects model was made for the mean differences of the proposed outcomes was performed (total sleep time, sleep onset latency and number of nighttime awakenings). Inter-study heterogeneity was evaluated with I2 parameter. RESULTS: The initial search yielded 355 results, of which three met the selection criteria. Melatonin was found to be a safe and effective drug to increase total sleep time and decrease sleep onset latency in children and adolescents with ASD; so far the evidence on the number of night awakenings is not statistically significant. DISCUSSION: According to available evidence, melatonin is a safe and effective choice for managing certain sleep problems in children and adolescents with ASD. Studies are needed with larger sample sizes and comparing with other drugs available in the market.