Objective: The purpose of this study was to develop a machine learning-based model (eXtreme Gradient boost [XGBoost]) that can accurately predict the severity of acute cholangitis in patients. The model was designed to simplify the classification process compared to conventional methods. Methods: We retrospectively collected data from patients with cholangitis who visited the emergency department of a secondary medical institution in Seongnam, Korea from January 1, 2015 to December 31, 2019. The patients were divided into three groups (Grade I, II, III) based on severity according to the Tokyo Guidelines 2018/2013 (TG18/13) severity assessment criteria for cholangitis. We used algorithms to select variables of high relevance associated with the grade of severity. For the XGBoost models, data were divided into a train set and a validation set by the random split method. The train set was trained in XGBoost models using only the top seven variables. The area under the receiver operating characteristic (AUROC) and the area under the precision-recall curve (AUPRC) were obtained from the validation set. Results: 796 patients were enrolled. The top 7 variables associated with the grade of severity were albumin, white blood cells, blood urea nitrogen, troponin T, platelets, creatinine, prothrombin time, and international normalized ratio. The AUROC values were 0.881 (Grade I), 0.836 (Grade II), and 0.932 (Grade III). The AUPRC values were 0.457 (Grade I), 0.820 (Grade II), and 0.880 (Grade III). Conclusion We believe that the developed XGBoost model is a useful tool for predicting the severity of acute cholangitis with high accuracy and fewer variables than the conventional severity classification method.