Abstract: Fuzzy cluster analysis theory is introduced, three methods of data standardization for cluster analysis are discussed and the extremity standardization method is recommended. This method was applied to the fault diagnosis of suction fan. The fault data and normal data were input to compose the sample data, the fuzzy equivalent matrix based on the transitive closure was imposed to the cluster analysis, and the checking data were introduced to validate this method. A practical example shows that the result calculated by fuzzy cluster analysis is in good agreement with the actual inspections, indicating that the method is available for the fault diagnosis of suction fan.