Abstract <p>Heart disease is a primary cause of disability and premature mortality globally. Coronary heart disease is the most prevalent kind of heart disease, which happens when plaque builds up inside the arteries that feed blood to the heart, making blood circulation difficult. Heart disease prediction is a difficult task in clinical machine learning. However, various existing systems are utilized to detect the type of heart disease but those approaches are time-consuming and inaccurate to detect the disease at early stages. To address various issues, a deep learning framework has been developed to achieve accurate disease classification. Initially, data’s are collected and pre-processed using a Sequential K-Nearest Neighbors (SKNN) technique for missing value replacement. The data is then subjected to decimal scaling normalization to enhance its integrity and uniformity. Then, reducing the dimension of the feature vector by applying Multilinear Principal Component Analysis (MPCA). Butterfly optimization (BOA) is employed to determine the ideal quantity of components to enhance the accuracy of the proposed model. In order to determine the different forms of cardiac disease, characteristics are classified subsequently using Long Short-Term Memory (LSTM). To evaluate the planned model’s performance, performance measures from the proposed and existing models are compared. Performance measures include Sensitivity, MCC, Negative Predictive Value (NPV), False Discovery Rate (FDR), Accuracy, Precision, Error, Specificity, F1-score, False Negative Rate (FNR), False Positive Rate (FPR), False Negative Rate (FNR), and False Positive Rate (FPR) attained for the proposed model is 96.5, 95, 3.5, 95.9, 95.5, 94.7, 95.7, 2.8, 3.7, 90.9, 93.2, 95.7 and 2.9%. In comparison to other existing techniques, the proposed technique performs better. In order to determine the type of heart disease, the created model is the best choice.</p>