Air pollution is a great concern to mankind and is causing too many adverse effects on every living organism on earth by increasing lung diseases, skin diseases, and many other problems caused by it. This research presents a comprehensive study on the application of heterogenous ensemble learning techniques for PM2.5 concentration prediction, aiming to enhance prediction accuracy and provide insights into the driving factors behind pollution levels. The primary objective is to conduct a comparative analysis of heterogenous ensemble method, namely, blending and stacking in conjunction with individual base models, such as multiple linear regression (LR), decision trees (DT), support vector regression (SVR) and artificial neural networks (ANN). In total 28 models were created using blending and 28 models were created using stacking. Hyperparameter tuning is done to optimize the models.