Industrial Control System (ICS) plays important role to reduce the human interact to operate the industrial system process. Cyber Physical Systems (CPSs) exist in critical infrastructure such as nuclear power generation, transportation networks, gas and water distribution networks, Unmanned Aerial Vehicle Systems (UASs) and electric power distribution networks. In this paper, we present a system to detect anomalies and malicious activities in critical water infrastructure. This system helps the industrial operator and administrator when an anomaly occurs and acts on the infrastructure.  The system is built using various machine learning techniques such as Logistic Regression (LR), Linear Discriminant Analysis (LDA), Classification and Regression Tree (CART) and Support Vector Machine (SVM). The model was evaluated using a real-world dataset covering 15 anomaly scenarios including normal system behavior. The presented scenarios covered a wide range of events, ranging from hardware failure to sabotage in the water critical infrastructure. The overall evaluation showed that CART is the best classification technique because it has the highest results in all performance evaluation metrics such as accuracy, precision. There is a comparative study between the results after applying normalization on the dataset. The results after applying normalization are better than the results before applying it.