Water quality is incredibly important in both environmental and financial aspects. Water bodies are vulnerable to contamination and hence quality of water is a serious issue. Therefore water quality prediction is crucial before using it for any purpose. Modeling and prediction of organic pollutants in water has become vital in water pollution control. This work utilizes regression algorithms and minimal remotely sensed parameters like pH, DO, Temperature, BOD, TSS, Ammonia to predict the Chemical Oxygen Demand (COD). The COD calculation model is established using weighted multiple linear regression and compared with other regression algorithms. From the test results of the advanced model, it is observed that the model using weighted multiple linear regression has excellent prediction efficiency, is less computationally expensive, uses minimal parameters and is of practical application value in IOT environment, thereby avoiding the use of costly sensors. It can be implemented at the edge and can also be used to monitor the water level changes.