Abstract: Objective： Our aim was to explore and evaluate factors that predict mortality in acute fatty liver of pregnancy (AFLP) patients and built a effective predictive model. Methods： This retrospective study included 93 AFLP patients admitted to Xiangya Hospital and the second Xiangya Hospital, Central South University， Hunan, China. Univariate and multivariate logistic regression analyses were used to identify predictors of mortality and built predictive model, while odds ratios (OR) and ROC curves were used to evaluate the importance of the risk factors. Results： The international normalized ratio (INR), total bilirubin, creatinine, platelet count, and hepatic encephalopathy were independently related to poor outcomes. The area under ROC curve of the model based on platelet count, and hepatic encephalopathy was 87.0%，with 81.8% sensitivity and 76.8% specificity. Conclusions： Elevated total bilirubin, INR, creatinine, low platelet counts, hepatic encephalopathy are independent significant variables that predict the mortality of AFLP patients. The predictive accuracy of the model for AFLP was 87.0%.