The fraudulent financial statement of a company is becoming serious over the last few years, so, finding a valid forecasting fraudulent financial statement model is an urgent work for academic research and financial practice. The UTilities Additives DIScriminants (UTADIS) classification method is an effective approach to classify some data into different groups. Therefore, based on UTADIS method, this paper uses the utility function theory and linear programming theory to classifying the financial data of a company as fraudulent or non fraudulent class, in which the research dataset of the company consists of 10 financial ratios. The empirical results show that UTADIS method can effectively detect the fraudulent financial data.