本研究旨在由傳統財務指標及公司治理角度分析近年來我國因財務報表舞弊遭一審判決有罪之公司發生的舞弊警訊(Red Flags)，建立財務報表舞弊預警模型，作為會計師執行舞弊偵測時的參考，降低因查核舞弊所產生的審計失敗，並可作為主管機關實質審查選案標準及投資人選擇投資標的之參考。 本文以民國85年至95年因財務報表舞弊遭法院一審判決有罪之27家公司為樣本，分別使用財務及公司治理變數建立財務報表舞弊預警模型，再以Cascade Logistic迴歸分析結合財務及公司治理模型，檢視Cascadev Logistic模型的正確判別率。實證結果顯示，財務變數中舞弊當期應付帳款成長率、舞弊當期存貨成長率、應收帳款收現天數對前期同一比率與財務報表舞弊呈顯著正相關；舞弊當期資產報酬率對前期同一比率、舞弊當期銷貨成長率與財務報表舞弊呈顯著負相關。公司治理變數中董監持股、控制持股、外部個人監事席次與財務報表舞弊顯著負相關；財報重編次數呈顯著正相關，良善的公司治理機制與財務舞弊具負向關連性。而整合的Cascade Logistic模型，財務及公司治理變數均與財務報表舞弊呈顯著相關，且均具增額解釋。 / Presented are a profile of a sample of fraud companies in Taiwan from 1996 to 2006, their financial and corporate governance characteristics, suggested fincial and corporate governance logistic models for detecting fraud and a final combine cascade logistic model. The results suggest both systematic relationships between the probability of fraud and financial and corporate governance characteristic variables. This evidence is consistent with the usefulness of accounting data in detecting fraud and the effectiveness of proper corporate governance in protecting fraud happening. Because the cascade logistic model correctly identifies approximately 90% of the companies involved or not invleved in fraud, the model can be a useful screening device for auditors to detect fraud, low down the propobility of audit failure and the cost of lawsuit due to fraud.