There are very few studies are carried for investigating the potential of hybrid ensemble machine learning techniques for building a model for the detection and prediction of heart disease in the human body. In this research, the authors deal with a classification problem that is a hybridization of fusion-based ensemble model with machine learning approaches, which produces a more trustworthy ensemble than the original ensemble model and outperforms previous heart disease prediction models. The proposed model is evaluated on the Cleveland heart disease dataset using six boosting techniques named XGBoost, AdaBoost, Gradient Boosting, LightGBM, CatBoost, and Histogram-Based Gradient Boosting. Hybridization produces superior results under consideration of classification algorithms. The remarkable accuracies of 96.51% for training and 93.37% for testing have been achieved by the Meta-XGBoost classifier.