Melatonin is effective for migraine prevention in adults. We hypothesized that melatonin would also be effective for migraine prevention in children and adolescents. This was a randomized, double-blind trial of melatonin (3 mg or 6 mg) versus placebo for migraine prevention in 10-17 year-olds with 4-28/28 headache days at baseline. Participants were recruited from the UCSF Child & Adolescent Headache Program, UCSF child neurology clinic, and social media advertisements. Migraine diagnosis was confirmed by a headache specialist. Participants completed an 8-week single-blind placebo run-in. Those meeting randomization criteria (≥4 headache days and ≥23/28 electronic diary entries during weeks 5-8) were randomized 1:1:1 to placebo:melatonin 3 mg:melatonin 6 mg nightly for 8 weeks. The primary outcome measure was migraine days in weeks 5-8 of randomized treatment between melatonin (combined 6 mg + 3 mg) versus placebo. We aimed to enroll n = 210. The study closed early due to slow enrollment (n = 72). Two participants were in the single-blind phase when the study closed, therefore the meaningful n = 70. Sixteen percent (11/70) were lost to follow-up during the single-blind phase. An additional 21% (15/70) did not meet randomization criteria (<4 headache days: n = 5, <23/28 diary days: n = 7, both: n = 3). Sixty-three percent (44/70) were eligible to randomize, of whom 42 randomized (n = 14 per arm). Taking another preventive at enrollment (OR 8.3, 95% CI 1.01 to 68.9) was the only variable associated with meeting randomization criteria. Of those randomized, 91% (38/42) provided diary data in the final 4-weeks. However, given the amount of missing data, only those with ≥21/28 diary days were analyzed-7/14 (50%) in the placebo group, and 20/28 (71%) in the melatonin groups combined. Median (IQR) migraine/migrainous days in weeks 5-8 of double-blind treatment was 2 (1-7) in the placebo group versus 2 (1-12) in the melatonin groups combined; the difference in medians (95% CI for the difference) was 0 days (-9 to 3). There were no differences in adverse events between groups. When compared to recall at enrollment, headache days decreased across the single-blind placebo phase and the double-blind phase. There was no suggestion of superiority of melatonin; however, given the substantial portion of missing data, numerically higher in the placebo arm, and underpowering, this should not be interpreted as proof of inefficacy. Melatonin was generally well tolerated with no serious adverse events. Future migraine preventive trials in this age group may find this trial helpful for anticipating enrollment needs if using a single-blind placebo run-in. Enriching for those already on a migraine preventive may improve randomization rates in future trials, though would change the generalizability of results.