To expand the breeding flock of Poll Dorset sheep in China, multiple ovulation and embryo transfer breeding program was applied to the limited number of imported Australian Poll Dorset sheep. This study investigated the effects of FSH from three different manufacturers, parity (nulliparous vs multiparous), repeated superovulation, oestrus induction, corpus luteum regression and oestrus delay on Poll Dorset superovulation. The results showed that gonadotropin FSH from Canada Folltropin-V (Ca-FSH) was successfully used for superovulatory treatment with 160 mg-200 mg dosage per ewe and recovered 12.91 ± 7.80 embryos. Multiparous ewes for superovulation treatment were significantly better nulliparous ewes (p<0.05). The successive superovalution treatment reduced significantly embryo collection but did not affect transferable embryo number. Ewes with natural oestrus resulted in significantly higher number of embryos (13.83 ± 4.64) and of transferable embryos (12.00 ± 5.76) than ewes with induced oestrus (7.00 ± 4.92; 4.22 ± 3.42) and unknown oestrus cycle (5.94 ± 3.38; 3.19 ± 2.56, p<0.05). The delayed oestrus ewes at 24 h after superovulatory treatment produced significantly fewer embryos and transferable embryos (0.92 ± 1.51 vs 0.42 ± 0.90) than those with normal oestrus (p<0.01). Furthermore, the more transferable embryos were recovered from ewes with normal corpus luteum than those with corpus luteum regression (5.88 ± 5.09 vs 3.59 ± 4.30 and 8.83 ± 5.75 vs. 6.66 ± 5.41, p<0.01). These results suggest that in our farm practice, a comprehensive treatment method by using the Canadian FSH (Folltropin-V), plus choosing multiparous and natural oestrus ewes with normal corpus luteum might obtain an optimum embryo collection and embryos transfer in sheep.