Heart disease is still one of the world’s top causes of mortality. Thus, prevention and effective treatment depend on early detection. This study uses the Cleveland Heart Disease Dataset to examine how ML techniques can be used in the prediction of heart disease. By removing outliers, encoding categorical data, handling missing values, and scaling features, the dataset was prepared for further processing. There was an 80:20 split between the data sets used for training and testing. Data was collected and used to train and assess a number of classification models. These models included DT, SVM, RF, and ANN. In comparison to the other models, the ANN performed quite well, achieving 86% accuracy, 86% precision,84% recall, and 83%F1- score. In contrast, DT, SVM, and RF showed lower performance across all metrics, with ANN proving to be the most reliable for heart disease prediction. The study concludes that ANN offers the highest predictive capability, making it a promising tool for early heart disease detection. Future research could explore the incorporation of additional features, such as lifestyle factors or genetic data, to enhance model accuracy.