The CNC system is a complex mechatronics system, which make it difficult to diagnose fault. Expert system for fault diagnosis that utilizes domain knowledge and the profiles of experts to fix the problem of the complex system has become an important issue. A hybrid expert fault system combining the rule-base reasoning (RBR) with case-based reasoning (CBR) for CNC system is proposed. The combination can trouble-shoot rapidly, improve the CNC system reliability and maintainability. The hybrid system is implemented by using QT and SQLITE database. The experiment result of the system shows that the system diagnosis efficiently and accurately.