Abstract: The theory of gray clustering was applied in condition assessment of canned motor pump in order to solve the problem for lack samples, low mensurability and bad reachable maintainability. Considering the small number of samples and fuzzy interval, each clustering index was whitened and quantified according to the partition of condition of each index and the weight was given in the light of the significance of each clustering index. The condition of canned motor pump was distinguished by applying the acquired quantified value of clustering index and weight. The results indicate that this method is objective, rational and fitting for the condition assessment of nuclear power equipment for lack samples such as canned motor pump.