To evaluate the impact of mifepristone-induced abortion (MA) on the duration of third stage labour in a subsequent pregnancy, an observational cohort study was conducted from 1998 to 2001 at antenatal clinics in Shanghai, Beijing and Chengdu, China. A total of 4925 pregnant women with no history of induced abortion (NA) and 4931 pregnant women with one previous MA were enrolled and followed until delivery. Of these, 5139 women who delivered singletons vaginally were used in the present analyses, including 2614 with NA and 2525 with a history of MA. Maternal characteristics, labour duration and other obstetric and gynaecological information were obtained. The incidence rates of prolonged third stage of labour were 1.55% and 1.49% in NA and MA, respectively. After adjusting for age at delivery, maternal education, maternal occupation, area of residence, duration of gestational, type of delivery and pregnancy-induced hypertension, MA was not associated with the risk of prolonged third stage of labour (odds ratios = 0.92, 95% confidence interval 0.58, 1.44). Subgroup analysis of women with MA showed similar results regardless of gestational age at abortion, woman's age at abortion, subsequent curettage/complications and the interpregnancy interval. In conclusion, the data did not provide evidence that one MA was associated with the risk of prolonged third stage of labour in a subsequent pregnancy in primiparae.