Objective To develop a model for predicting feeding intolerance risk in children with congenital heart disease (CHD) after cardiac surgery, and to provide reference for identifying children at high risk. Methods A retrospective cohort study of 265 children younger than 1 year with CHD was conducted. Clinical data were collected for model training, and seven machine learning algorithms were performed including logistic regression, support vector machine, random forest, decision tree, gradient boosting decision tree, extreme gradient boosting, and plain Bayesian. Accuracy, precision, recall, F-score and the area under the receiver operating characteristic curve (AUC) were used to evaluate the model performance. Results The incidence of feeding intolerance in children with CHD after cardiac surgery was 49.4%. Among the seven machine learning algorithms, the performance of the extreme gradient boosting was the best, with an AUC of 0.914 (95%CI:0.849, 0.967). Ten factors including vasoactive-inotropic score, duration of mechanical ventilation before feeding, feeding initiation time, age, etc.affected feeding intolerance in children with CHD. Conclusion Feeding intolerance is relatively high in children with CHD after cardiac surgery. The prediction model developed based on the extreme gradient boosting demonstrates good performance, which may contribute to clinical decision-making and optimization of individualized enteral nutrition management.