Objective To compare the effectiveness of natural cycles (NC) and artificial cycles (AC) in women undergoing frozen-thawed embryo transfer (FET) after in vitro fertilization (IVF). Design Retrospective cohort study. Setting Public fertility center in China. Population We studied 9733 women undergoing the first cycle of FET over a 3-year period (June 2014 - December 2017). All women were followed-up until one year after embryo transfer. Methods The type of endometrial preparation was determined by the treating physician’s preference, based on patients’ characteristics. Women with regular ovulation were allocated to natural cycles (n=1676), while patients who were reluctant to frequently monitoring or living far from the hospital were allocated to artificial cycles (n=8057). A logistic regression model was used to assess the association between endometrial preparation and clinical outcomes while adjusting for potential confounders. Main outcome measures Live-birth rate was primary outcome while miscarriage rate, clinical pregnancy rate, preterm birth rate, and ectopic pregnancy rate were secondary outcomes. Results In the adjusted model, type of endometrial preparation did not affect live birth (OR, 0.89; 95%CI, 0.79-1.01), clinical pregnancy (OR, 0.96; 95%CI, 0.85-1.09), preterm birth (OR, 1.09; 95%CI, 0.90-1.33) and ectopic pregnancy (OR, 0.77; 95%CI, 0.36-1.61), while AC significantly increased the miscarriage rate (OR, 1.38; 95%CI, 1.11-1.73, P=0.004). Conclusion In women undergoing FET, natural cycles and artificial cycles resulted in comparable live birth rate while miscarriage rate was higher in artificial cycles.