Abstract: To improve on the data mining and utilize the efficiency of enterprise resource planning(ERP),and enhance the function of prediction and decision support of ERP systems,a method is proposed to learn some decision action in ERP systems through the improved decision tree algorithm.Based on the learned result,some prediction support for the new decisions can be obtained in ERP systems.The corresponding approach of algorithms is demonstrated,the test results verify that the improved decision tree algorithm is effective and feasible in predicting decision supports.