Coronary Artery Disease (CAD) is the broad spectrum of heart-related illnesses ranks among the most prevalent conditions affecting middle-aged individuals cause of mortality worldwide, necessitating accurate and discovering the disease at early stages allows healthcare providers to save the patient's life. In this study, we have evaluated the performance of five machine learning algorithms (ML): Random Forest Classifier (RFC), Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), and Naive Bayes (NB), for the classification of CAD. The dataset was subjected to k-fold cross-validation with varying values of k (5, 7, 10, and 15) to assess the robustness and consistency of the models. The performance metrics that used to evaluate performance of the algorithms was compared were confusion matrices and average accuracy scores as the evaluation metrics. Our results revealed significant differences in the predictive capabilities of the models across different folds, highlighting the impact of cross validation on model reliability. The findings provide insights into CAD prediction by the most suitable machine learning approaches 4 and underscore the importance of model evaluation strategies in medical data analysis.