Comorbid cocaine abuse adversely affects clinical outcomes in schizophrenia. Using a prospective, randomized, parallel group design (N = 24), we tested the hypothesis that patients with schizophrenia treated with olanzapine have reduced cocaine craving and abuse compared with those treated with haloperidol. In addition, we examined whether this differential effect correlated with reductions in extrapyramidal symptoms, positive and negative symptoms, and/or depression. There were no significant differences overall in proportions of positive drug screens between treatment groups; no differences in positive, negative, or depressive symptoms; and few differences between treatment conditions in extrapyramidal symptoms. However, craving for cocaine was rated significantly lower by patients treated with haloperidol compared with patients treated with olanzapine. Important study limitations include a small sample size and high attrition rates. Larger controlled studies are necessary to determine optimal antipsychotic therapy for patients with schizophrenia and comorbid cocaine abuse.