Longitudinal effects of child maltreatment on cortisol regulation in infants from age 1 to 3 years were investigated in the context of a randomized preventive intervention trial. Thirteen-month-old infants from maltreating families (N = 91) and their mothers were randomly assigned to one of three intervention conditions: child-parent psychotherapy, psychoeducational parenting intervention, and a control group involving standard community services (CS). A fourth group of infants from nonmaltreating families (N = 52) and their mothers comprised a nonmaltreated comparison (NC) group. The two active interventions were combined into one maltreated intervention (MI) group for statistical analyses. Saliva samples were obtained from children at 10:00 a.m. before beginning a laboratory observation session with their mothers when the children were 13 months of age (preintervention), 19 months (midintervention), 26 months (postintervention), and 38 months (1-year postintervention follow-up). At the initial assessment, no significant differences among groups in morning cortisol were observed. Latent growth curve analyses examined trajectories of cortisol regulation over time. Beginning at midintervention, divergence was found among the groups. Whereas the MI group remained indistinguishable from the NC group across time, the CS group progressively evinced lower levels of morning cortisol, statistically differing from the MI and NC groups. Results highlight the value of psychosocial interventions for early child maltreatment in normalizing biological regulatory processes.