Abstract: As a knowledge representation of fuzzy uncertainty, fuzzy conceptual graphs have attracted wide attention from scholars of artificial intelligence. In MORTON's representation of fuzzy conceptual graphs, fussiness has certain limitation. To overcome this limitation, we combine fuzzy mathematics theory with traditional fuzzy conceptual graph and study fuzzy conceptual graph interms of fuzzy uncertain knowledge representation. An improved representation of fuzzy conceptual graph is presented in this paper. Fuzzy set is used to represent the fuzziness of concepts and conceptual relations in fuzzy conceptual graph. The storage structure of fuzzy conceptual graph is designed according to this representation, and the matching operation rules of fuzzy conceptual graph are established. Improved fuzzy conceptual graph has been applied to methods of automatic checking over subjective examination, the accuracy is high.