Water is indispensable for the development and maintenance of all forms of life. There are major issues with the availability and quality of drinking water on a global scale. Drinking water contaminated with pathogenic germs, hazardous compounds, etc., can be dangerous to one's health. In this study, we describe a procedure for evaluating water excellence and providing early warning of impending contamination. The water might be contaminated by a wide variety of factors. These factors are considered and used to calculate the optimal frequency of water purification. Internet of Things (IoT) and Machine Learning are utilised by the system. It includes a physical and chemical sensor to check the parameters (pH, conductivity, turbidity, temperature, and humidity of the water in the tank). The information collected by the sensors are uploaded to a database and then analysed. Predictions are made using the Sequential Learning Neural Network (SLNN) method. There are a lot of factors to be considered when setting up a neural network, and finding the optimal setup may be a tedious and time-consuming process. The Sequential Learning Neural Network method is used for this task because it is efficient and may significantly reduce the amount of time and information storage required. It creates a non-linear relationship between input and output. When any of a user's parameters fall below predefined thresholds, the system issues a warning message. The client can get an early warning concerning water contamination in their storage tanks. This method may be used in large-scale water treatment plants as well as in private homes with storage tanks.