Abstract: In order to effectively solve the influence of the actual machining position of the tool on the radial error of the machine tool spindle, a radial thermal error modeling and prediction method based on tool deflection is proposed. The method takes the spindle of vertical machining center HNC715 as the research object, and uses the spindle analyzer to collect the radial thermal error of the spindle. After analyzing the data of the radial thermal error of the spindle of the machine tool, linear regression analysis is carried out, and different speeds are constructed. The radial thermal error model of the upper and lower ends of the lower rod is detected. By analyzing the tool deflection principle and the radial thermal error model of the upper and lower ends of the detection rod, a comprehensive modeling method of the radial thermal error of the machine tool spindle is proposed, and the method is verified by experiments. The verification results show that the radial thermal error comprehensive modeling method of the machine tool spindle is effective and the prediction accuracy is high.