Cardiovascular disease is the reason for global death nowadays is a crucial disease, since it takes away the lives of children, youngsters, and aged people. Early prediction is mandatory to avert the loss and many researchers had effectuated heart disease prediction models. The prediction accuracy must be high to provide effective treatment. In context with this, we propose a novel stacking approach for the IoT-based prediction of heart diseases with the utilization of K-Nearest Neighbors (KNN), Naïve Bayes (NB), Extreme Gradient Boosting (XGBoost), and Feed –Forward Neural network (FNN). The parameters such as Blood pressure, SPO2, and ECG are collected from the respective sensors and stored in cloud storage for further processing. The collected data are pre-processed using Min-Max Normalization to delete the unnecessary and blank data. Simulations are conducted in Python and validate the performances of the proposed approach with state-of-art works such as OCI-DBN, SMOTE, CNN, AND EDCNN with the metrics such as log-loss, specificity, F1-score, ROC, accuracy, Matthews’ correlation coefficient and precision.