Integrated care for children is rarely studied, especially in low- and middle-income countries, where generalists often provide mental health care. To explore the effect of adding a child and youth component to an existing adult collaborative care program on mental health outcomes and receipt of care. This cluster randomized trial was conducted within an adult collaborative care program in Tehran, Iran. General practitioners (GPs), their 5- to 15-year-old patients, and patients' parents were included. Children and youths coming for routine medical visits who scored greater than the cutoff on the Strengths and Difficulties Questionnaire (SDQ) were followed up for 6 months. The study was conducted from May 2018 to October 2019, and analysis was conducted from March 2020 to August 2021. GPs were randomized to either a 2.5-day training on managing common child mental health problems (intervention) or refresher training on identification and referral (control). Primary outcome was change in SDQ total problems score; secondary outcomes included discussion of psychosocial issues by the GPs and receipt of mental health care during the follow-up period. Overall, 49 GPs cared for 389 children who scored greater than the cutoff on the SDQ (216 children in intervention group, 173 in control group). Patients' mean (SD) age was 8.9 (2.9) years (range, 5 to 15 years), and 182 (47%) were female patients. At 6 months, children in the intervention group had greater odds of receiving mental health care during the study (odds ratio [OR], 3.0; 95% CI, 1.1 to 7.7), parents were more likely to report that intervention GPs had discussed parent (OR, 2.1; 95% Cl, 1.1 to 3.8) and child (OR, 2.0; 95% Cl, 0.9 to 4.8) psychosocial issues, and intervention GPs were more likely to say they had provided counseling (OR, 1.8; 95% Cl, 1.02 to 3.3). However, there was no greater improvement in SDQ scores among children seen by intervention vs control GPs. Adjusted for clustering within GP, the variables used for balanced allocation (practice size, practice ownership, and study wave), and the other variables associated with change in SDQ scores over time, there was not a significant time-treatment interaction at either the 3- or 6-month follow-up points (linear combination of coefficients for intervention, 0.57 [95% CI, -1.07 to 2.22] and -0.08 [95%CI, -1.76 to 1.56], respectively). In a subgroup of GPs with practices composed of 50% or more children, children seen by intervention GPs improved to a significantly greater extent (-3.6 points; 95% CI, -6.7 to -0.46 points; effect size d = 0.66; 95% CI, 0.30 to 1.01) compared with those seen by control GPs. In this cluster randomized trial, GP training on managing common child mental health problems did not demonstrate greater improvement in child SDQ scores. Child mental health training for GPs in collaborative care can improve children's access to mental health care, but prior experience working with children and their families may be required for GPs to use a brief training in a way that improves child outcomes. ClinicalTrials.gov Identifier: NCT03144739.