Support Vector Machine (SVM) is a new machine learning method based on statistical learning theory. This method has obvious advantages in processing small-sample and nonlinear problems. As a matter of fact, the generation of earthquake is a very complicated nonlinear dynamic problem and the earthquake data manifest the nonlinear and irregular characteristics. This paper analyzed the seismic precursor of Tianjin City and neighborhood systematically and presented a synthetic earthquake predication model with a method of SVM classification by employing the seismic precursor information which reflects the short-term situation of 2~3 months. The results show that the method is effective and has a good application future.