As you know, the uterus is part of the reproductive system of women and is. When normal cells in the cervix turn into abnormal cells and gradually grow abnormally out of control, they become cervical cancer (cervical cancer).The aim is to compare the several initiatives to better accuracy, the foresight for the cancer to develop from various western dominated methods by using the engine to identify the strong features of the coast's network CNN-SVM. It is the integration The abstract features extracted from the hidden layer LeNet-5 are combined with strong features after dimension reduction, then the combined features are entered into the SVM classifier for classification. Finally, they are classified by SVM, KNN, ANN, DROPUT, and by voting method they are we also compared. The results showed that the strong feature of the CNN-SVM model can be used in cell classification for early cervical cancer screening. In the conclusion obtained, as well as using functions to improve performance, such as logistic regression, with Acc, Sn, Sp 93.6, 00, 00 relatively, only performed more than in the part and well, and it could have performed more than for improvement.