PURPOSE: Epilepsy, one severe prevalent brain disorder, primarily relies on drug treatment. However, approximately one-third of patients with epilepsy do not achieve effective control with current medications, underscoring the need for more innovative treatment approaches. Notably, melatonin has gained attention for its anti-seizure properties and favourable safety profile. This systematic review aimed to evaluate the efficacy and safety of melatonin as an add-on treatment for epilepsy. METHODS: We searched for articles published before June 2023 in Web of Science, Cochrane Library, and PubMed. We used RevMan 5.4 software to compute relative risks (RRs) and 95 % confidence intervals (CIs). Key outcomes included total sleep time, wakefulness after sleep onset, sleep latency, seizure frequency, seizure severity, and safety. The quality of randomised controlled studies (RCTs) was assessed using the Cochrane Risk of Bias tool. RESULTS: Of the 264 publications retrieved, 10 RCTs were included in the meta-analysis. Add-on melatonin treatment improved sleep latency (RR: 0.56; 95 %CI: 0.10-1.02; P = 0.02) and seizure severity (RR: 0.33; 95 %CI: 0.04-0.62; P = 0.03) compared with placebo treatment. Adverse events (increased headache severity in children with a history of migraines, bronchitis, ear infections, agitation, and urinary frequency) were reported in only one trial. CONCLUSION: This systematic review found that add-on melatonin therapy improved sleep latency and seizure severity in patients with epilepsy. However, several of the included studies did not systematically assess sleep quality, seizures, and safety and lacked long-term follow-up data. Further RCTs with extended follow-up periods are required to definitively determine the efficacy and safety of melatonin.