Abstract This study presents the anticipation precision of the modeling thermal deformation. It measures the machine temperature and thermal deformation of the surface grinder. This study uses two methods of regression analysis:Moving-Average、Auto Regression Moving-Average, and two methods of Neural Network:Feed-forward 、Recurrent, to find the relationship equation between the machine temperature and thermal deformation. Using the way of simulation tests and verifies the anticipation precision of the modeling thermal deformation established by four methods. It expects to find the best anticipation model of thermal deformation.