<p>Coronary heart disease (CHD) is a massive health problem globally because it contributes to higher rates of morbidity and mortality. Early and specific prediction of cardiac disease is necessary for successful prevention and control. Machine learning (ML) methods can be utilized to predict cardiac disease based on medical reports, allowing individuals to conveniently assess their condition from home. Based on a dataset of medical records and clinical attributes, ten ML models were evaluated namely: decision tree, K-nearest neighbors, gradient boosting, random forest, ADABoost, support vector machine, logistic regression, naive Bayes, hypertuned gradient boosting, StackingCV ensemble. The models were assessed by evaluating performance metrics such as accuracy, precision, sensitivity, specificity, and F1 score. The random forest classifier accomplished an accuracy of 0.91, demonstrating its high discriminatory power in identifying cases of CHD. Intriguingly, the hypertuned gradient boosting model significantly outperformed the baseline model, achieving an impressive accuracy of 0.93. Finally, the StackingCV ensemble model outperformed all others, achieving the highest accuracy, sensitivity, specificity of 0.95, along with superior precision, and F1 score, making it the most robust model for CHD prediction. To further enhance model transparency, explainable AI techniques such as SHAP and LIME were employed to identify the most influential features and provide interpretability of model predictions. These findings highlight the potential of ML models, particularly ensemble methods, in the early detection of coronary heart disease and support their integration into clinical decision-support systems.</p>