Abstract: By using the concepts and methods of fuzzy logic, the fuzzy premise and matching of the fuzzy rules in the knowledge base of FMS fault diagnosis and maintenance system (FMSFDMS) developed by the authors were defined and explained. The GMT fuzzy logic reasoning method employed by FMSFDMS was presented, and a way for determining the uncertainty measure membership function of the hypothesis finding was also discussed. For the convenience of human machine interaction, the uncertainty measure membership values of the finding were divided by the multiple thresholds, and then expressed by the linguistic variables.