This study examines the effects of treated and untreated depressive symptoms on the likelihood of utilization of highly active antiretroviral therapy (HAART) among a multi-site cohort of HIV-infected women who screened positive for probable depression. Data were collected biannually from 1996 through 2001 in a prospective cohort study. Random-effects regression analysis was used to estimate the longitudinal effects of mental health treatment on the probability of HAART utilization, controlling for clinical indicators (CD4 count, viral load), demographic features (race/ethnicity, income), and behavioural factors (recent crack, cocaine, or heroin use). Use of antidepressants plus mental health therapy, or use of mental health therapy alone significantly increased the probability of HAART utilization, compared to receiving no depression treatment. Use of antidepressants alone did not differ significantly from receiving no depression treatment. African American women and those who used crack, cocaine, or heroin also were less likely to use HAART. These findings suggest that efforts to enhance depressed women's access to psychopharmacologic treatment and therapy may increase their use of the most effective HIV therapies.