Prediction of heart disease at early stages is considered to be critical to manage and control it. Several techniques in machine learning (ML) and data mining can be employed to predict the presence of heart disease. The main goal of this research is to develop a model based on traditional ML and data mining such as Decision Trees, Support Vector Machines, Artificial Neural networks and Bayesian classifiers in order to predict the presence of heart disease and to assist physicians in maximizing accuracy for identification of the heart disease severity stage. In addition, we will also employ deep learning techniques (such as CNN, DNN, and RNN), Transfer Learning (TL), Federated Learning (FL), and compare their performance with the traditional ML techniques. In this research, we will use the database sets mentioned below for more than a thousand of instances that include important clinical, physical, laboratory and historical patient data: Coronary heart disease historical data | IEEE DataPort (ieee-dataport.org) Heart Disease Dataset (Comprehensive) | IEEE DataPort (ieee-dataport.org)