Prednimustine is an ester of chlorambucil and prednisolone. The drug has had some activity in experimental as well as in human tumors, including breast cancer. The objective of the current study is to compare the clinical effectiveness and toxicity of prednimustine with its components of chlorambucil plus prednisolone, and to compare the clinical effectiveness and toxicity of a continuous treatment with an intermittent one. Two hundred and one patients with advanced breast cancer were randomized to the following treatments: I: prednimustine continuously; 40 mg/m2 d; II: prednimustine intermittently, 160 mg/m2 d 1 to 5, repeated every three weeks; and III: chlorambucil 8 mg/m2 d plus prednisolone 8 mg/m2 d, both given continuously. All treatments were given orally. One hundred and ninety five patients were evaluable for assessment of response to the treatments. The response rates were 21% in both of the prednimustine‐treated groups, whereas only 11% of the patients treated with chlorambucil and prednisolone responded. The duration of remission with continuous prednimustine was longer than with the other two treatment groups, but not significantly so. One hundred and ninety five patients were evaluable for assessment of hematologic toxicity. No differences were observed between the two prednimustine‐treated groups, whereas the group treated with chlorambucil plus prednisolone experienced a significantly higher degree of leukopenia and/or thrombopenia. Subjective toxicities, nausea, vomiting and psychical symptoms were only moderate in the three treatment groups, but significantly more pronounced in the group treated with intermittent prednimustine. Because of these promising results, prednimustine should be compared to other alkylating agents in patients with advanced breast cancer in randomized trials.