OBJECTIVE To evaluate the role of recytoreductive surgery (RCRS) in recurrent epithelial ovarian cancer. METHODS Fifty-five patients who underwent recytoreductive surgery at Peking Union Medical College Hospital between Nov. 1998 and Apr. 2002 were retrospectively reviewed. They were divided into four groups firstly by the identity of recurrent lesion: group A, single lesion; B, disseminated lesion; C, intestinal obstruction; D, palliative surgery. Then they were divided into 3 groups according to the sensitivity to chemotherapy. Group A, diseases relapsed more than 6 months after chemotherapy; group B, relapsed less than 6 months; group C, relapsed during chemotherapy. We review the cases with regard to its macroscopic residual disease, complications of operation, cases with complete remission and partial remission, postoperation survival time and disease-free interval. RESULTS Recytoreductive surgery for patients with isolated recurrent tumor were optimal. For those diagnosed as local lesion preoperatively, 61% of these cases were found to have disseminated diseases postioeratively. Sixty-five percents of these cases received optimal RCRS, but their prognosis were not as well as those with real isolated leasion (response rate: 36% vs 67%). Cases with different sensitivity to chemotherapy had different prognosis (response rate: A, 50%; B, 26%; C, 28%). CONCLUSIONS Recytoreductive surgery should be considered in patients with isolated recurrent ovarian cancer and patient with recurrence more than 6 months after chemotherapy.