Purpose/Significance To construct a risk prediction model for job burnout among nurses in tertiary hospitals in China, and to explore the optimal algorithm. Method/Process By using the questionnaire survey method, 23 risk factor data are collected from nurses in tertiary hospitals in Liaoning province, and their job burnout status is evaluated through the Copenhagen burnout inventory. Seven single algorithms, namely random forest, extra-trees, gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), K-nearest neighbors (KNN), support vector machine (SVM), and logistic regression, and a stacking ensemble model strategy are used to construct risk prediction models. Model performance is evaluated using indicators such as AUC, accuracy, and specificity.Result/Conclusion Among the single models, logistic regression performs the best. Among the stacking ensemble models, the model integrating random forest and logistic regression is the optimal one. The ensemble model can serve as an efficient tool for the screening of nurses’ job burnout.