Bariatric surgery has become a powerful therapy for patients with obesity and type 2 diabetes mellitus (T2DM) whereby significant weight decreases are achieved, and metabolic remission can take place. However, remission rates differ widely, making predictive tools that support personalized care crucial.The intent of the present study was to generate and test machine learning (ML) and deep learning (DL) models for predicting T2DM remission following bariatric surgery in addition to detect major clinical and behavioral predictors through explainable AI approaches.Cross-sectional study at King Fahad Specialist Hospital from January to April 2024. Clinical, anthropometric, and behavioral data were used to train classification models (10 ML and 4 DL).SHapley Additive exPlanations (SHAP) were employed to interpret the models. Analysis of performance metrics including ROC AUC, accuracy, precision, recall, and F1-score was also performed.Bottleneck Network showed the highest ROC AUC of 0.889, followed by Random Forest (0.863). SHAP analysis revealed the percentage of weight regain, BMI, A1c levels, and dietary habits as significant predictors.Explainable AI models provide a logical and interpretable model to predict T2DM remission after bariatric surgery for the personalized decision-making and personalized care of patients.   Keywords Bariatric surgery, Hypoglycemia, machine learning