Abstract:This paper puts forward the principle of SVM classification based on Principal Components Analysis, and apply it to the research in Financing Analysis. The main idea is to normalize the attributes of the sample for training and testing and to calculate the degree of correlation of different attribute at first, then to build up new attribute sets and apply it to SVM for classify. This paper has tested the two patterns classification for 106 companies listed on China Stock Exchange by 2000 and the result indicates that it can improve the accuracy on the recognition of model.