Coronary Artery Disease (CAD) is a major cause of morbidity and mortality worldwide, highlighting the need for accurate and non-invasive prediction methods. Machine learning techniques have shown promise in supporting CAD diagnosis using structured clinical and electrocardiographic data. In this study, a machine learning–based framework is investigated for CAD prediction using the integrated heart disease dataset. Multiple feature selection methods and classification models are evaluated to analyze their impact on predictive performance. Experimental results indicate that although feature selection identifies clinically relevant attributes, it does not consistently improve classification accuracy across all models. In contrast, a hybrid voting ensemble approach demonstrates more stable and reliable performance by effectively integrating multiple classifiers. These findings highlight the advantage of ensemble learning for CAD prediction and support its potential use in clinical decision-support systems.