Abstract. The incidence of premature corpora lutea (CL) regression was investigated in three different breeds (Merino x polled Dorset: MD, Suffolk: S, and South-Down: SD), parous and nonparous ewes (a total of 46 ewes) treated for superovulation during and out of the breeding season. Superovulation was induced with a single injection of 20 mg porcine follicle-stimulating hormone (pFSH) 2 days and 500 IU equine chorionic gonadotropin (eCG) 1 day, respectively before the removal of synthetic progestogen (40 mg FGA) vaginal sponges inserted for 12 days. All ewes were inseminated into the uterus with fresh-diluted or frozen-thawed semen by laparoscope, and a number of newly formed and regressed CL were recorded. Breed difference significantly (P<0.005) affected the incidence of premature regressed CL; S (6.3%, n=16) ewes were less in the proportion of the incidence of regressed CL than MD (37.5%, n=16: P<0.05) and SD (71.4%, N=14, P<0.01) ewes. The use of non-parous ewes and treatment during autumn tended to result in higher rates of incidence of premature CL regression than the use of parous ewes and treatment during spring. These results indicate that the incidene of premature CL regression in superovulated ewes was influenced by not only season and parity but also by breed of ewes used for superovulation treatment.