A total of 334 HIV-infected patients with oral candidiasis were randomly assigned to receive 14 days of treatment with either 100 mg of oral fluconazole once daily or 10 mg clotrimazole five times daily. Both treatments were clinically effective: 98% of evaluable fluconazole-treated patients and 94% of evaluable clotrimazole-treated patients were cured or showed improvement (p = NS). Fluconazole was more effective than clotrimazole in eradicating Candida from the oral flora by the end of therapy (65% versus 48%) (p = 0.005). In addition, patients in the fluconazole-treated group were more likely to remain asymptomatic through the second week of follow-up (82.3% versus 50.0%) (p < 0.001). This difference was no longer evident by the post-therapy visit during week 4. Seven patients treated with clotrimazole and two patients treated with fluconazole discontinued therapy because of side effects. Two patients in the fluconazole group were withdrawn from therapy because of elevated serum glutamic-oxaloacetic transaminase levels, one considered possibly related to drug therapy. Fluconazole was as effective as clotrimazole in the treatment of oral candidiasis and temporarily provided a more prolonged disease-free state. Future studies are needed to define the optimal regimen for both the treatment and prevention of recurrent oral candidiasis in HIV-infected patients, addressing special attention to the issue of compliance, cost, and emergence of resistance.