This study aimed to systematically review and synthesise the available evidence on the prevalence and associations between psychotic-like experiences (PLEs) and substance use in children and adolescents aged ⩽17 years, prior to the typical age of development of prodromal symptoms of psychosis. As substance use has been associated with earlier age of psychosis onset and more severe illness, identifying risk processes in the premorbid phase of the illness may offer opportunities to prevent the development of prodromal symptoms and psychotic illness. MEDLINE, PsycINFO, and CINAHL databases were searched for chart review, case-control, cohort, twin, and cross-sectional studies. Study reporting was assessed using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist, and pooled evidence was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. Searches identified 55 studies that met inclusion criteria. Around two-in-five substance users reported PLEs [rate = 0.41, 95% confidence interval (CI) 0.32–0.51; low quality evidence], and one-in-five with PLEs reported using substances (rate = 0.19, 95% CI 0.12–0.28; moderate-to-high quality evidence). Substance users were nearly twice as likely to report PLEs than non-users [odds ratio (OR) 1.77, 95% CI 1.55–2.02; moderate quality evidence], and those with PLEs were twice as likely to use substances than those not reporting PLEs (OR 1.93, 95% CI 1.55–2.41; very low quality evidence). Younger age was associated with greater odds of PLEs in substance users compared to non-users. Young substance users may represent a subclinical at-risk group for psychosis. Developing early detection and intervention for both substance use and PLEs may reduce long-term adverse outcomes.