The primary goal of the current study was to examine the impact of maternal criticism (expressed emotion-criticism; EE-Crit) on the prospective development of depressive episodes in children. In addition to examining baseline levels of EE-Crit, we also sought to determine whether distinct subgroups (latent classes) of mothers could be identified based on the levels of EE-Crit they exhibited over a multiwave assessment and whether that latent class membership would predict depression onset in children. Finally, we examined whether EE-Crit and maternal depression would independently predict children's depression risk or whether EE-Crit would moderate the link between maternal depression and children's depression onset. Children of mothers with or without a history of major depression (N = 100) were assessed 5 times over 20 months. Children completed the Children's Depression Inventory and mothers completed the Five Minute Speech Sample and the Beck Depression Inventory at the baseline assessment, and at 2-, 4-, and 6-month follow-up assessments. Children and mothers completed diagnostic interviews assessing children's onsets of depressive episodes at the 20-month follow-up. Latent class analysis of the 4 waves of EE-Crit assessments revealed two distinct groups, exhibiting relatively lower versus higher levels of EE-Crit across the first 6 months of follow-up. EE-Crit latent class membership predicted children's depression onset over the subsequent 14 months. This finding was maintained after controlling for mother's and children's depressive symptoms during the initial 6 months of follow-up. Finally, maternal depression did not moderate the link between EE-Crit and childhood depression onset. Continued exposure to maternal criticism appears to be an important risk factor for depression in children, risk that is at least partially independent of the risk conveyed by maternal depression. These results highlight the importance of a modifiable risk factor for depression—repeated exposure to maternal criticism.