This paper is carried on the heuristic prediction of rainfall using machine learning techniques. As we know agriculture was the predominant of our country and economy. While a regular rain pattern is usually played vital for healthy agriculture but too much rainfall or too little rainfall can be harmful, even it led to devastating of crops. This paper discusses the rate of rainfall in previous years according to various crops seasons like rabi, Kharif, zaid and predicts the rainfall in future seasons. The paper also measures the different categories of data by linear regression method in metrics for effective understanding of agriculture in India. We have selected a real dataset which consists of past year's rainfall rate according to various seasons. Results of this application help farmers to make a correct decision to harvest a particular crop accordingly to crops seasons. Linear regression helps to find. © 2017 IEEE.