Sleep problems are common in children, especially those with neurodevelopmental disorders, and can lead to consequences in behavior, functioning, and quality of life. We systematically reviewed the efficacy and harms of pharmacologic treatments for sleep disorders in children and adolescents. We searched MEDLINE, Cochrane library databases, and PsycINFO through June 2018. We included 22 placebo-controlled randomized controlled trials (1-13 weeks' duration), involving 1758 children (mean age 8.2 years). Single randomized controlled trials of zolpidem and eszopiclone in children with attention-deficit/hyperactivity disorder (ADHD) showed no improvement in sleep or ADHD ratings. Clinical Global Impression Improvement/Severity scores significantly improved with zolpidem ( P = .03 and P = .006, respectively). A single, small randomized controlled trial of diphenhydramine reported small improvements in sleep outcomes (8-10 minutes' better sleep latency and duration) after 1 week. In 19 randomized controlled trials, melatonin significantly improved sleep latency (median 28 minutes; range 11-51 minutes), sleep duration (median 33 minutes; range 14-68 minutes), and wake time after sleep onset (range 12-43 minutes), but not number of awakenings per night (range 0-2.7). Function and behavior improvement varied. Improvement in sleep was greatest in children with autism or other neurodevelopmental disorders, and smaller in adolescents and children with chronic delayed sleep onset. Adverse events were infrequent with melatonin, but more frequent than placebo in children taking eszopiclone or zolpidem. These findings show that melatonin was useful in improving some sleep outcomes in the short term, particularly those with comorbid ASD and neurodevelopmental disorders. Other drugs and outcomes are inadequately studied.