Financial restatement has significant effect on financial report users, capital market and restating businesses. The purpose of this study is to develop a warning model of the financial restatement to forecast the risk factors inducing the restatement. This paper uses logistic model and data mining techniques- neural networks and decision trees to analyze data and compare accuracy and error rate and construct the optional financial restatement prediction model. The paper hopes to build the financial restatement prediction model to give the reference to publics.