Purpose Mitoxantrone-corticosteroid is currently the standard palliative treatment in hormone-refractory prostate cancer (HRPC) patients. Recent clinical trials documented the high activity of the docetaxel-estramustine combination. We conducted a randomized phase II study to evaluate prostate-specific antigen (PSA) response (primary end point) and safety of two docetaxel-estramustine-prednisone (DEP) regimens and mitoxantrone-prednisone (MP). Patients and Methods One hundred thirty metastatic HRPC patients were randomly assigned to receive docetaxel (70 mg/m2 on day 2 or 35 mg/m2 on days 2 and 9 of each 21-day cycle) and estramustine (280 mg PO tid on days 1 through 5 and 8 through 12) or mitoxantrone 12 mg/m2 every 3 weeks; all patients received prednisone (10 mg daily). Results One hundred twenty-seven patients were assessable for PSA response and safety. A ≥ 50% PSA decline was found in a greater percentage of patients in the docetaxel arms (67% and 63%) compared with MP (18%; P = .0001). Median time to PSA progression was five times longer with DEP than with MP (8.8 and 9.3 v 1.7 months, respectively; P = .000001). Overall survival was better in the docetaxel arms (18.6 and 18.4 months) compared with the MP arm (13.4 months), but not significantly so (P = .3). Crossover rates differed significantly among treatment arms (16%, 10%, and 48% in arms A, B, and C, respectively; P = .00001). Treatment-related toxicities were mild and mainly hematologic. Conclusion The results of this randomized phase II study showed significantly higher PSA decline ≤ 50% and longer times to progression in HRPC patients receiving DEP-based chemotherapy than MP, and that DEP could be proposed in this setting.