Data were collected from 30 companies on the Vietnam Stock Exchange from 2016 to 2020, this study attempts to uncover factors that influence financial restatement and establish rules based on data mining techniques to forecast future financial restatement. WEKA statistical software is used in conjunction with the Cross-Industry Standard Process for Data Mining (CRISP-DM) to preprocess and combine various datasets gathered for data analysis. The research findings are two-fold. First, point out that a high likelihood of a financial restatement may be linked to specific features. The existence of fraudulent financial statements has a significant relationship with five main variables, namely Accounts Receivable ratio, Accounts Payable, accounts payable to sales, Debt to Equity, and Evidence of CEO Changed. Second, this study provides a number of rules to predict future financial restatement, helping investors to predict the restatement status of the financial statement’s items, thereby assessing the company’s fraud tendency before making an investment decision. Our findings help investors obtain better evaluations of fraud probability that might occur by recognizing the main attributes associated with financial restatements. Moreover, investors can use certain rules to assess the fraud tendency based on companies’ future financial restatement, which might enable them to dig more into financial fraud signals before finalizing their investment. The uniqueness of our study also lies in the employment of data mining techniques in the field of financial accounting, which allows our study to provide better insight into financial fraud detection and prediction.