In order to improve the recognition speed of cervical cell and obtain the highest recognition accuracy with the least number of features，this paper innovatively uses the Classification and Regression Trees（CART） algorithm to select features，and then the Particle Swarm Optimization（PSO） algorithm is used to optimize the Support Vector Machine(SVM). Therefore，the PSO-SVM classification algorithm is formed to classify the cells. This paper uses the Herlev dataset to verify the validity of the proposed algorithm. Through the CART feature selection method，9 representative features are successfully extracted from 20 features，and the accuracy of two classifications and seven classifications are above 99%. Further，this paper introduces several other classification and recognition algorithms of cervical cancer cells for simulation comparison. It can be founds that the recognition accuracy of this algorithm is obviously superior when the number of features is small，which indicates that the proposed algorithm is effective. The method effectively reduces the difficulty of artificial feature selection，and ensures that the recognition accuracy of the cells is almost the same as before when the recognition time is reduced. Thus，the proposed algorithm provides an effective method for the diagnosis of cervical cancer diseases.