Now a days, in this competitive world, every organization has been investing their major resources like money and manpower towards predicting the customers behavior in maximizing their profitability. Understanding the public opinion in advance to their product release, helps the organizations in taking effective decisions on marketing strategies, Churn Prediction is an approach, through which one can easily attain the behavior of customer as Life Time Value. In our research, we focus more on building a model, that can train the input data records of about 2,666 having 21 attributes of different type and deploy the constructed rules of about 105 to generate reports on churn prediction using Decision tree algorithm. The records in the Dataset, considered in our experiment are classified as positive (2280) and negative (386) manually. In our research, we found that there is good correlation score(0.82) exist between Dates delay and days to settle attributes, upon implementing our proposed model, we achieved 85% of classification accuracy with 70.1% cohen's kappa value. Finally, we would like to conclude that much higher classification accuracy can be achieved using Decision tree algorithm by constructing a very low level Decision rules on the data with good domain knowledge.