To determine the factors predicting the probability of severe postpartum hemorrhage (SPPH) in women undergoing repeat cesarean delivery (RCD). This multicenter, retrospective cohort study involved women who underwent RCD from January 2017 to December 2017, in 11 public tertiary hospitals within 7 provinces of China. The all-variables model and the multivariable logistic regression model (pre-operative, operative and simple model) were developed to estimate the probability of SPPH in development data and external validated in validation data. Discrimination and calibration were evaluated and clinical impact was determined by decision curve analysis. The study consisted of 11,074 women undergoing RCD. 278 (2.5%) women experienced SPPH. The pre-operative simple model including 9 pre-operative features, the operative simple model including 4 pre-operative and 2 intraoperative features and simple model including only 4 closely related pre-operative features showed AUC 0.888, 0.864 and 0.858 in development data and 0.921, 0.928 and 0.925 in validation data, respectively. Nomograms were developed based on predictive models for SPPH. Predictive tools based on clinical characteristics can be used to estimate the probability of SPPH in patients undergoing RCD and help to allow better preparation and management of these patients by using a multidisciplinary approach of cesarean delivery for obstetrician.