Cardiovascular complications are a major concern among individuals with diabetes, requiring reliable and interpretable risk-prediction tools. Machine learning offers strong predictive capabilities, but many models lack transparency, limiting their clinical usefulness. Current risk-prediction systems often function as “black boxes,” making it difficult for clinicians to understand why a model flags a patient as high risk. There is a need for an explainable and accurate framework that identifies key risk factors and supports personalized decision-making for diabetic patients. This study proposes an explainable machine learning framework for predicting cardiovascular risk in diabetic patients using data from the Behavioral Risk Factor Surveillance System (BRFSS 2015). The model integrates the CatBoost classifier with SHapley Additive exPlanations (SHAP) to provide both accurate predictions and trans-parent explanations. After selecting records of diabetic patients, the model was trained on 22 health and lifestyle features. It achieved 69.3% accuracy, AUC = 0.738, and F1-score = 0.482. SHAP analysis identified age, sex, difficulty walking, high blood pressure, and education as key contributors to cardiovascular risk. The framework also generated individual-level SHAP force plots, explaining each patient’s risk profile in a simple and interpretable way. This approach supports personalized diabetes management by combining data-driven prediction with clinical explainability. Future work will include model validation on external datasets and integration into clinical decision-support tools.