In this paper, we solve problem of the ineffectiveness of using RDFS/OWL-stored mechanism for large-scale domain ontology. In particular, when the constructed ontology contains a huge amount of entities and semantic relations, it causes difficulties in managing and visualizing the ontological knowledge as well as low performance in data querying. We resolve these issues by using graph database as the storage mechanism for representing the constructed ontology. The approach of using graph database provides the advantages not only in better ontological data management and visualization but also in the higher performance and flexible of knowledge extracting from ontology via cypher querying language.