Purpose/Significance To identify the key influencing factors of type 2 diabetes mellitus (T2DM) with hypertension, and to provide a basis for risk prediction and personalized intervention in high-risk populations. Method/Process Data of 3 839 inpatients from the department of endocrinology from 2020 to 2022 are extracted. Four analysis models of comorbidity influencing factors, namely random forest, support vector machine (SVM), extreme gradient boosting (XGBoost), and natural gradient boosting (NGBoost), are constructed and compared. The data of 1 000 inpatients in 2023 are used as the validation set. The model performance is verified by combining SHAP analysis and age-stratified subgroup validation. Result/Conclusion The NGBoost model shows the optimal performance in predicting T2DM with hypertension. SHAP analysis reveals that smoking, number of complications and thyroid diseases are key influencing factors. Additionally, the model exhibits good stability across four age subgroups.