In order to enhance security of the mobile phone, a method of user operation behavior authentication based on SVM is proposed. By monitoring the mobile phone touch screen device, continuous access to user operation when the sliding track, contact area and other raw data. The user behavior feature extraction algorithm is designed to establish the user characteristic sample, and the SVM algorithm is used to train the user behavior characteristic model. The user access goals and historical certification results are integrated with different authentication strategies to achieve key protection sensitive data to facilitate the user access to non-sensitive data. Experimental verification in the Android system environment shows that the method has a good authentication effect.