Water is the most vital resource for living things, especially for human beings. In order to be healthy, everyone must have taken good-quality water in his/her entire life. However, checking the quality of drinking water is challenging task. The traditional way of checking drinking water quality methods involves manual sampling of waters followed by laboratory analysis. This process may be difficult to cover a large geographical area and needs a huge amount of budget for running this process. This research aims to monitor the quality of drinking water based on wireless sensor network (WSN) and machine learning algorithms. We have used water temperature, PH and turbidity sensors to collect real-time data from different water contents. These sensors sense different parameter of waters and transmit the sensed data to the base station using Wi-Fi module. For detecting and classifying the abnormal behavior of the drinking water, we used k nearest neighbor (KNN) supervised machine learning classification algorithm on the collected data. The reason of using KNN classification is that its appropriateness for having small amount of data sets and suitability for having more than two class datasets. Evaluation is performed on k nearest neighbor classification algorithms using accuracy as evaluation metrics. In this research the KNN classifiers classifies the data with the accuracy of 94%. The results show that the KNN algorithm is having better performance on wireless sensor data with multi labeled datasets compared to SVM classifier.