Abstract: Taking manpower resource management problems of the Ministry of Metallurgical Industry as background, the concrete design method of Manpower Resource Intelligent Management System (MRIMS) is given. In this system, general management analysing and forecasting model tree is used as its knowledge representation. Its system structure is supported by multi-base structure. In its inference engine, meta-knowledge inference is applied to coordinate with quantitative inference, qualitative inference and neural network inference. Finite natural language understanding is introduced into the man-machine interface of the system. TURBO PROLOG and C languages are used to implement the design partly on IBM/AT microcomputer.