The school-based Playing-2-gether is a 12-week intervention with two components aimed at decreasing child externalizing behavior through improving teacher-child interactions. The first component is rooted in attachment theory and aimed at enhancing teacher-child relationship quality, and the second is based on learning theory and aimed at improving teachers’ behavior management. In this three-wave randomized study, effects of Playing-2-gether on the teacher-child relationship quality and on teacher-rated child behavioral adjustment were investigated. To this aim, 175 dyads consisting of male preschoolers with relatively high levels of externalizing problem behavior and their teachers were randomly assigned to Playing-2-gether (n = 89) or an education-as-usual control condition (n = 86). Teacher-rated questionnaires were collected at pre-test, after the first intervention component, and at post-test. At post-test, the intervention group showed a larger decrease in teacher-child conflict, child conduct problems, and child hyperactivity/inattention. Supplementary analyses showed that all positive effects were already visible after the first intervention component and that teacher-child conflict, child conduct problems and hyperactivity/inattention did not further reduce during the second component. In addition, an increase in closeness was found following the first component, but subsequently disappeared at post-test.