Numerous illnesses that affect the heart are categorized as heart diseases, commonly referred to as cardiovascular diseases. These can include heart rhythm issues, blood vessel issues, etc., related to congenital cardiac abnormalities. In recent years, it has been the leading cause of death worldwide. Thus, it can save many people’s arteries if it can detect this dangerous issue early. Accurate and reliable methods must be found quickly to automate the process of early disease diagnosis and achieve effective disease management. In addition, it has been concentrated on an effective strategy to raise the effectiveness. In the previously existing systems, the heart disease dataset is applied to K-Nearest Neighbors (KNN), Naive Bayes, and Convolutional Neural Network (CNN) for the prediction. The proposed system uses the info-gain feature selection technique to eliminate unneeded features and extract the required features. One of the supervised machine learning algorithms called random forest is used to forecast the presence of cardiac disease in patient bodies. The optimization technique called hyperparameter tuning is used to tune some hyperparameters in the random forest algorithm to eliminate overfitting. The UCI (University of California, Irvine) heart disease dataset obtained from the Kaggle repository is used. Precision, recall, and F1-Score are used to calculate the accuracy. The optimized algorithm is then used to train and test the data, thus predicting the presence of heart disease in people, and increased accuracy is obtained.