Direct marketing is rapidly growing over the past few years within a variety of industries, especially insurance, banking and other financial services among for obtaining an immediate action. Data mining models are completely helping those organizations to set marketing goals and improve the likelihood of response. The objective of this work to build the prediction model with high performance which examines bank customer behavior using realworld data, whether a client will subscribe a term deposit or not which dataset collected Portuguese Bank marketing campaign. The paper investigates Multiple Adaptive Regression Splines using the multi-stage feature selection method. Besides, it contributes to enhancing the campaign effectiveness and important decision in the business area.