Concurrent abuse of cocaine and heroin is a common problem. Methadone is effective for opioid dependence. The question arises as to whether combining agonist-like or antagonist-like medication for cocaine with methadone for opioid dependence might be efficacious. Two parallel studies were conducted. One examined sustained release d-amphetamine and the other risperidone for cocaine dependence, each in combination with methadone. In total, 240 subjects (120/study) were recruited, who were both cocaine and heroin dependent and not currently receiving medication. All provided consent. Both studies were carried out for 26 weeks, randomized, double-blind and placebo controlled. Study I compared sustained release d-amphetamine (escalating 15-30 or 30-60 mg) and placebo. Study II examined risperidone (2 or 4 mg) and placebo. All subjects underwent methadone induction and were stabilized at 1.1 mg/kg. Subjects attended clinic twice/week, provided urine samples, obtained medication take-home doses for intervening days, and completed self-report measures. Each had one behavioral therapy session/week. In Study I, reduction in cocaine use was significant for the 30/60 mg dose compared to the 15/30 mg and placebo. Opioid use was reduced in all groups with a trend toward greater reduction in the 30/60 mg d-amphetamine group. In Study II, methadone reduced illicit opioid use but cocaine use did not change in the risperidone or placebo groups. There were no adverse medication interactions in either study. The results provide support for the agonist-like (d-amphetamine) model in cocaine dependence treatment but not for antagonist-like (risperidone) treatment. They coincide with our previous reports of amphetamine or risperidone administered singly in cocaine-dependent individuals.