Background Our indicated Prevention program for preschool children with Externalizing Problem behavior (PEP) demonstrated improved parenting and child problem behavior in a randomized controlled efficacy trial and in a study with an effectiveness design. The aim of the present analysis of data from the randomized controlled trial was to identify the mediating processes that account for these positive treatment effects. We hypothesized that a decrease in negative parenting and increases in positive parenting and parental warmth would mediate the relationship between treatment and child improvement. Method Parents of 155 children were randomly assigned to the PEP intervention group (n = 91) or a nontreated control group (n = 64). Parents rated their child's problem behavior and their own parenting practices before and after PEP training. Parental warmth was assessed during standardized play situations. Four mediation models were tested using structural equation modeling. Trial registration number ISRCTN12686222; http://www.controlled-trials.com/isrctn/pf/12686222. Results Changes in child externalizing problem behavior were most strongly mediated by reductions of negative parenting in difficult parenting situations. Increases in positive parenting also served as a mediator. Changes in parental warmth, parents' feeling of self‐efficacy, and parental mental health did not play a mediating role in the association between PEP treatment and child behavior. Conclusions In our program, the most important component was to teach parents how to reduce dysfunctional parenting strategies in conflict situations.