Service churn often occurs in business and management processes due to the quality changes of product/service, competition environment, and other reasons. Service churn prediction has an important position in the strategic planning of each company and organization. In this study, we propose a prediction model based on ensemble learning to address this problem. To build the model, six conventional machine learning models and one artificial neural network with multilayer perceptron are utilized for base models of the ensemble learning. With the stacking technique, the resulting model achieved the best performance compared with other relevant models. In addition, integrating an imbalanced data processing method is also helpful to enhance the overall performance of the proposed method.