Customer churn is an important concern of many industries and business organizations. It is an active area of research since it is directly related to the revenue of the business. Early prediction of customer churn helps the industries to take preventive action to retain their customer. There are a lot of machine learning techniques applied in this area and they differ in computational cost and accuracy. In this paper, we propose stochastic gradient boosting to predict customer churn and compare the results with the existing techniques. We have tested the proposed method with two different data set namely Telecom and Banking. For these two data set our approach achieved better accuracy as well as recall rate when compared to other machine learning algorithms. Copyright © 2018 American Scientific Publishers All rights reserved.