Based on financial management and enterprises of early-warning theory, this paper constructs a financial distress early-warning model using GA-SVM. First, it uses listed companies appearing in Shanghai Stock Exchange and Shenzhen Stock Exchange in 2007-2009 as sample books. Defining ST listed companies which have abnormity of finance status as signature of the listed company's financial crisis. Then it uses the data in the financial statements known to the public as the input feature vector and combine genetic algorithm and support vector machine. Use Taking an empirical research with the financial distress early-warning model. Test results of the demonstration study shows the model has a superiority in predicting financial distress.