Reasonable selection of technological parameters plays an important role on the CNC grind machining effect on engineering ceramics for the caver machine. But the relationship between technological parameters and machining effect is extremely complex and it is very difficult to build the relational model by traditional regression method. In order to solve this problem, a BP neural network prediction model of CNC grind machining of engineering ceramics is built on the basis of grind machining characteristics by using neural network theory. Simulation and experimental results prove the validity of the prediction model. The prediction model can be used to reasonably select the technological parameters for CNC grind machining of engineering ceramics and improve the machining quality and machining efficiency.