Abstract: the final performance of CNC machine tools was directly determined by the quality of the assembly, in order to predict the assembly quality of CNC machine tools beforehand and improve the assembly qualification rate, the GA-SVM method is proposed to predict the assembly quality prediction model of CNC machine tools.First, CNC machine tools was decomposed into meta-actions based on"Function-Motion- Action" functional structural decomposition, the influential factors in meta-actions were regraded as the assembly influential factors, and the kinematics parameters of the terminal meta-action in meta-action chain were taken as the assembly quality analysis object. Then, the GV-SVM model was applied to assembly quality prediction of the meta-action chain under the X-axis feed motion of grinder wheel frame. To demonstrate the practicality and effectiveness of the proposed method, the prediction results obtained by the GV-SVM model were compared with those obtained by the BP neural network and the conventional SVM model, and the results show that the average relative errors of the three prediction results are 3.83%, 8.90% and 10.16%, respectively. Obviously, the GA-SVM model has the best prediction effect than the other two prediction models, the proposed model provides theoretical guidance for optimizing the assembly process of CNC machine tools.