CNC equipment is the high cost of the products, it cost too much to get the failure time of CNC equipment from life test alone. In order to evaluate the reliability of CNC machine tools is very important to study the performance of the numerical control equipment and the time-dependent pattern when the experimental data is not very rich. The hidden semi- Markov model ( HSMM ) was put forward for data mining on CNC equipment performance parameters. The state transition matrix was calculated which describing the failure process that CNC equipment from the normal state to the failure of the state transition. Combined state transition matrix and Markov prediction model to assess the reliability of CNC equipment. The feasibility of this method is verified in a numerical control milling machine, this method enriches the CNC equipment reliability assessment theory.