Background: Parenting practices predict early childhood physical aggression. Preventive interventions that alter parenting practices and aggression during early childhood provide the opportunity to test causal models of early childhood psychopathology. Although there have been several informative preventive intervention studies that test mediation models in older children, no such studies have been conducted with younger children at high risk for psychopathology. Method: Within the context of a randomized controlled trial, we examined whether changes in parenting practices mediate the effects of a family intervention on observed physical aggression among African American and Latino younger siblings of adjudicated youths. Results: Improved parenting practices partially mediated the intervention effect on physical aggression. Improvements in harsh parenting, responsive parenting, and stimulating parenting explained a significant amount of the intervention effect on child physical aggression observed in the context of parent-child interactions. Parenting practices accounted for 38% of the intervention effect on physical aggression. Conclusions: There was support for the hypothesized model of the prevention of physical aggression during early childhood. Intervention benefits on parenting practices partially accounted for intervention effects on physical aggression in young high-risk children.