Root-cause identification and product quality characteristic modeling are key issues for improving product quality and the productivity of manufacturing process. The quality characteristic model, which indicates the influence of the dominate root-causes on the product quality characteristic, is the basis for quality control. In this paper an approach for quality characteristic modeling based on Support Vector Machine is presented. The selections of parameters and the Kernel function are discussed. The presented approach is applied for analyzing and predicting the quality of inertial gyroscope.