Nocturnal enuresis (NE) is common in school-age children with frequent bedwetting during the night. Previous studies have suggested that the children with enuresis may have abnormal Diffusion Kurtosis Imaging (DKI) characteristics. Thus, this study aims to propose an automated classification model by integrating DKI technology with machine learning algorithms to distinguish children with NE from healthy controls. A total of 93 children aged 6–12 years (47 with NE and 46 healthy controls) participated in the study. A range of imaging features were derived from DKI data. Classification analysis was conducted using Random Forest (RF), Gradient Boosting Decision Tree (GBDT), XGBoost, and ensemble models. The experimental results revealed that the ensemble model (RF-GBDT) achieved the best performance on the test set, with an accuracy of 85.7% and an AUC of 91.3%, significantly outperforming individual models (RF reached a maximum accuracy of 84.2% and AUC of 86.9%). The effectiveness of DKI features in classifying enuresis is validated, providing an objective and efficient auxiliary diagnostic tool for clinical applications.