This study investigated perinatal depressive symptoms among HIV-infected women enrolled in a cluster-randomized, controlled trial in South Africa. Women (n = 1370) attending 12 community health centers were consecutively enrolled in a two-phase (phase 1 = without a male partner, phase 2 = with a male partner) and two-condition (experimental or control) prevention of mother-to-child transmission (PMTCT) intervention. Women were enrolled at 8-24 weeks pregnant and followed postpartum at 6 weeks and 6 and 12 months (retention rate = 69.8%). Antenatally, 45.4% of women were above the 12-point Edinburgh Postnatal Depression Scale (EPDS) cutoff, 30.2% were above the cutoff at 6 weeks, and 34.2% and 36.9% at 6 months and 12 months postpartum, respectively. In multilevel regression analyses, depressive symptoms decreased over time among women in phase 2 participating in the intervention condition, but neither condition nor phase alone was associated with a decrease in depression. Greater HIV stigma, increased psychological intimate partner violence, less male involvement, lower education, and non-adherence during pregnancy were associated with increased depressive symptoms over the perinatal period. Results indicated that women participating had high levels of depressive symptoms (> 40% prenatally and > 30% postnatally), and the combination of the multi-session PMTCT intervention plus male partner participation contributed to a reduction in depressive symptoms. Results suggest that interventions targeting the reduction of depressive symptoms in perinatal HIV-positive women by increasing male involvement and decreasing HIV stigma and intimate partner violence are needed to reduce depression in this vulnerable population.