Background: The aim of this study was to confirm and compare the efficacy of fluvoxamine (the only licensed SSRI for treatment for OCD in Japan) and behavior therapy in treating Japanese patients with OCD. In addition, we investigated predictors of these treatments. Methods: Thirty-one outpatients meeting the DSM-III-R criteria for OCD without any axis I disorder were randomly assigned to one of three treatment conditions: BT (behavior therapy ± pill placebo), FLV [autogenic training (a psychological placebo for OCD) ± fluvoxamine] and control group [autogenic training (psychological placebo) ± pill placebo] for 12 weeks of treatment. The Yale-Brown Obsessive-Compulsive Scale (Y-BOCS) and the Clinical Global Impression-Improvement Scale (CGI-I) were administered blindly at baseline and week 4, 8 and 12. Results: Twenty-eight patients completed this study. Patients in the BT and FLV groups showed significantly more improvement than those in the control group in the mean score of total Y-BOCS; moreover, the BT group showed significantly more reduction in total Y-BOCS score at the end of treatment than the FLV group (BT > FLV, p < 0.01). Patients with lower baseline total Y-BOCS, past history of a major depressive episode and absence of cleaning compulsion improved more with fluvoxamine. Conclusions: We confirmed the effectiveness of behavior therapy and fluvoxamine for Japanese patients with OCD. Behavior therapy improved the condition of OCD patients more than fluvoxamine.