It requires staff to have high expertise to deal with faulty quickly during using and maintaining the CNC machine. This paper constructs the expert system of CNC machine faulty diagnosis and verifies prototype system based on CBR reasoning technology. According to research the characteristics of CNC machine faulty diagnosis to carding faulty structure and construct CNC machine faulty case base. Setting weight of different faulty attribute based on CNC machine characteristics. Due to the similar coefficient sum to compare problems to cases which in the faulty case base to select the most similar one as faulty diagnosis case in order to realize the reasoning of CNC machine faulty diagnosis depends .