Objective: Homework in cognitive–behavioral therapy (CBT) provides opportunities to practice skills. In prior studies, homework adherence was associated with improved outcome across a variety of disorders. Few studies have examined whether the relationship between homework adherence and outcome is maintained after treatment end or is independent of treatment attendance. Method: This study combined data from 4 randomized clinical trials of CBT for cocaine dependence to examine relationships among homework adherence, participant variables, and cocaine use outcomes during treatment and at follow-up. The data set included only participants who attended at least 2 CBT sessions to allow for assignment and return of homework (N = 158). Results: Participants returned slightly less than half (41.1%) of assigned homework. Longitudinal random effects regression suggested a greater reduction in cocaine use during treatment and through 12-month follow-up for participants who completed half or more of assigned homework (3-way interaction), F(2, 910.69) = 4.28, p = .01. In multiple linear regression, the percentage of homework adherence was associated with greater number of cocaine-negative urine toxicology screens during treatment, even when accounting for baseline cocaine use frequency and treatment attendance; at 3 months follow-up, multiple logistic regression indicated homework adherence was associated with cocaine-negative urine toxicology screen, controlling for baseline cocaine use and treatment attendance. Conclusions: These results extend findings from prior studies regarding the importance of homework adherence by demonstrating associations among homework and cocaine use outcomes during treatment and up to 12 months after, independent of treatment attendance and baseline cocaine use severity.