In this paper, we focus on the prediction and analysis of biogenetic data with high complexity by building integrated SVM models. Considering the complexity and high dimension of data set, we adopt the integration method based on sample segmentation to build the model. The results of the CCLE data analysis show that the model we used has better prediction results and smaller prediction variance than the generalized linear model, the integrated generalized linear model, and the original SVM model.