Diabetes, a chronic health condition with increasing prevalence, poses a significant challenge in the Philippines. This study proposes a machine learning web application that aims to predict an individual’s risk of developing diabetes based on the Philippine-specific lifestyle from the National Nutrition Survey (NNS) data warehouse. The model utilizes Bayesian-optimized TabNet architecture together with traditional supervised machine learning algorithms. To further improve the model’s predictions, the system integrates SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) as explainable artificial intelligence techniques. It aims to provide a reliable platform for predicting an individual’s risk of diabetes and act as a datadriven decision support tool.