上市公司财务舞弊长期存在并且危害甚大。本文从沪深两市上市公司中选取1994—2005年间126家财务报告舞弊公司和126家正常公司为样本,并选取29个指标作为解释变量试图建立财务舞弊定量识别模型。回归结果表明盈利能力弱、管理层持股比例高、独立董事人数少以及没有出具标准无保留意见的上市公司舞弊的可能性更大,通过Logistic回归技术得到的模型总体识别正确率达95．1%,预测效果显著。 The financial fraud of listed companies is long-term and harmful. This paper selects 126 financial reporting fraud companies and 126 normal companies from 1994-2005 listed companies in Shanghai and Shenzhen stock markets as samples, and selects 29 indicators as explanatory variables to attempt to establish the financial fraud quantitative identification model. The regression results show that the profitability is weak, the management holds a high proportion of shares, the number of independent directors is less and there is no possibility of issuing standard unqualified listed companies more likely to be fraud. The overall recognition accuracy of the model obtained by Logistic regression is 95.1 %, The prediction effect is remarkable.