In order to solve the problems of low fault monitoring efficiency,high maintenance cost and high delay of CNC machine tools under the condition of information technology,and meet the needs of intelligent diagnosis combining network technology,edge computing,fault prediction and health management technology，an intelligent diagnosis system based on the cyber physical system was proposed.The system was designed with four layers:intelligent sensing layer,data decision layer,network layer and application layer.Compared with the traditional fault diagnosis and health management,edge computing technology and artificial intelligence related technology were introduced to the system to solve the problems of real-time system,network reliability,data security and so on,and further to realize intelligent fault prediction and health management.Taking the ball screw pair of CNC machine tool as an example,by analyzing its fault phenomenon and designing prognostics and health management (PHM) process,the PHM algorithm was deployed and configured remotely by network technology,so as to realize online monitoring of the machine tool,and identify the early,middle and late faults of the ball screw.The results show that the system is of great significance to further improve the reliability of fault diagnosis of CNC machine tools and realize the intelligent operation and maintenance and health management of equipment.