In conventional water supply systems, continuous monitoring of water quality and the timely detection of leaks were challenging tasks for safeguarding public health and preserving water resources. They faced limitations in terms of scalability, real-time data processing, and energy efficiency. This research work aims to address the above challenges by developing a real-time prototype for monitoring the quality of water and leak detection. A novel voting-based ensemble model integrating correlation coefficient, mutual information, and Lasso regression for feature selection is proposed. IoT sensors are used for monitoring the quality of water and detecting leaks in the distributed water supply systems. The effectiveness of our approach is validated using the performance metrics accuracy, precision, recall, ROC-AUC curve, and F1 score. The proposed Ensemble Voting Feature Selection Approach outperforms with an accuracy of 93% and an MSE of 0.126.