The deadliest disease among humans is the heart disease, and it’s prevalent among all the ages of the people. Therefore, there is an urgent need for early detection of the heart disease, for both prevention and therapy purpose. However, the clinical methods used to diagnose HD today are expensive and frequently need for a high degree of intervention. Recently, varieties of intelligent systems for the automated diagnosis of HD have been designed by researchers as a solution to this problem. But, still there is a huge research gap in terms of accurate forecasting. This fact serves as the impetus for the current study's introduction of a new optimized Support Vector Machine (SVM) model for HD prediction. Pre-processing, feature extraction, feature selection, and heart disease prediction are the four main stages of the projected model. The initial pre-processing of the obtained data uses the Antonyan Vardan Transform (AVT) method. The features are then retrieved from the pre-processed data using the Interclass sub-Space Clustering method. The necessary characteristics are then retrieved using the Partially Differential Equation (PDE) from the extracted features. Finally, the heart disease is accurately predicted using the new SVM model that has been improved. The SVM is trained with new Enhanced Bat Optimization (EBO). The optimized SVM yields the final predicted result.