Objective This systematic review and meta-analysis aimed to (1) determine the accuracies of a broad range of screening tools for ADHD in children and adolescents and (2) compare the diagnostic accuracy of tools between population-based and clinical/high-risk samples, and across reporters. Method MEDLINE, PsycINFO, EMBASE and PubMed were searched up until February 20th, 2020 with no language restrictions. Studies reporting diagnostic accuracy of a screening tool against a diagnosis of ADHD in children <18 years were eligible for inclusion. Meta-analyses were undertaken to provide pooled estimates of the area under the curve (AUC), and sensitivity and specificity of groups of measures. Results Seventy-five studies published between 1985 and 2021 reporting on 41 screening tools that were grouped into four categories (ASEBA, DSM-IV symptom scales, SDQ, and Other Scales) were retained. The pooled AUC for studies using a combined ADHD symptoms score was 0.82 (95% CI 0.78-0.86), although this varied considerably across reporters (0.67-0.92) and populations (0.60-0.95). None of the measures met minimal standards for acceptable sensitivity (0.8) and specificity (0.8). Conclusion Most tools have excellent overall diagnostic accuracy as indicated by the AUC. However, a single measure, completed by a single reporter is unlikely to have sufficient sensitivity and specificity for clinical use or population screening.