<p>Proper selection of the key temperature measuring points (KTMPs) is an essential prerequisite for precise modeling of thermal errors for machine tools. A novel method for selecting temperature measuring points based on digital twin is proposed in this paper. Firstly, a spatial transmission model of the thermal error is constructed based on the transfer mechanism of thermal deformation. The transmission model demonstrated that the thermal deformations of all components continuously accumulate. So, the screening method for KTMPs based on the effect of thermal deformations of components (ETDC) is proposed. Then, a digital twin system for the thermal behavior of the machine tool is developed, which realizes the real-time acquisition of the thermal twin data and automatic screening of the KTMPs. Finally, The results show that the temperature nodes can be searched comprehensively by the twin data, and the established model can predict the thermal error more accurately. The KTMPs selected by the ETDC show good robustness to changes in modeling objects and operating conditions, which can significantly improve the accuracy.</p>