Objective: To investigate the factors associated with mental health service use among young children. Method: Five hundred ten preschool children aged 2 through 5 years were enrolled through 68 primary care physicians, with 388 (76% of the original sample) participating in a second wave of data collection, 12 to 40 months later. Consensus DSM-III-R diagnoses were assigned using best-estimate procedures. The test battery included the Child Behavior Checklist, a developmental evaluation, the Rochester Adaptive Etehavior Inventory, and a videotaped play session (preschool children) or structured interviews (older children). At wave 2, mothers completed a survey of mental health services their child had received. Results: In logistic regression models, older children, children with a wave 1 DSM-III-R diagnosis, children with more total behavior problems and family conflict, and children receiving a pediatric referral were more likely to receive mental health services. Among children with a DSM-III-R diagnosis, more mental health services were received by children who were older, white, more impaired, experiencing more family conflict, and referred by a pediatrician. Conclusions: Young children with more impairment and family conflict are more likely to enter into treatment. Services among young children of different races with diagnoses are not equally distributed. Pediatric referral is an important predictor of service use.