Maternal depression is prevalent, and puts children at risk. Little evidence addresses whether treatment for maternal depression is sufficient to improve child outcomes. An experiment was conducted testing whether psychotherapeutic treatment for mothers, suffering from major depression in the postpartum period, would result in improved parenting and child outcomes. Participants included depressed women randomly assigned to interpersonal psychotherapy (n = 60) or to a waitlist (n = 60), and a nondepressed comparison group (n = 56). At 6 months, depressed mothers were less responsive to their infants, experienced more parenting stress, and viewed their infants more negatively than did nondepressed mothers. Treatment affected only parenting stress, which improved significantly but was still higher than that for nondepressed mothers. Eighteen months later, treated depressed mothers still rated their children lower in attachment security, higher in behavior problems, and more negative in temperament than nondepressed mothers. Initial response to treatment did not predict reduced risk for poor child outcomes. Early maternal negative perceptions of the child predicted negative temperament and behavior problems 18 months after treatment. Treatment for depression in the postpartum period should target the mother-infant relationship in addition to the mothers' depressive symptoms.