We examined the differences in conflict interaction between depressed mothers and their toddler and non-depressed dyads and whether these differences mediated the association of maternal depression with compromised child socioemotional development. Mother/child interaction was videotaped during a teaching task and during a free play task as part of a home visit when the target child was between 16 and 18 months old. Each turn of every conflict was coded for behavior and affect of each member of the dyad. Interaction data were summarized to calculate the number of conflict turns, the rate of conflict, and the proportion of mother-initiated versus child-initiated conflicts per dyad. Sequential analysis was used to estimate the probability of specific maternal responses to specific child behaviors. Bivariate comparisons indicated that depressed dyads experienced higher rates of conflict, especially during the teaching task, and that depressed mothers were more likely to respond destructively to child oppositional behavior. Results of multivariate linear regression indicated that the higher probability of destructive response mediated the association of maternal depression with lower quality of mother-child attachment. These findings have implications for the development of interventions to support mothers in dealing with the conflicts that are so common during the second year of a child's life.