Access to clean and reliable water is a fundamental human right, yet ensuring a steady supply of safe and potable water can be a daunting task, particularly in urban areas where water distribution systems (WDS) are often complex and aging. However, despite of efforts to maintain and upgrade these systems, problems such as leaks, bursts, and contamination still persist, leading to significant economic, environmental, and health consequences. Conventional leak detection methods in WDS include acoustic, visual, and statistical approaches. However, these methods are often time-consuming and expensive and may not detect small leaks or those located in hard-to-reach areas. These limitations emphasize the need for more advanced and reliable leak detection techniques, such as those based on machine learning (ML). The objective of this study is to compare the performance of different ML models such as KNN Classifier, Random Forest Classifier, Support Vector Machine for leakage analysisconsidering no leak and leak scenarios. EPANET Example 3 is considered for hydraulic modeling using EPANET. Extended period simulation is performed and pressure and discharge relationship is considered. Accuracy score using the three techniques, The KNN Classifier, Random Forest Classifier, and Support Vector Machine are employed to evaluate the ML model’s performance which provided the accuracy score of 87.93%, 88.81% and 88.19%. It is observed that the use of Machine Learning based approaches supersedes the conventional methods in terms of efficient handling of large dataset in less time.