[Purpose/significance] The paper provides operational guidance for the association analysis of oral history archives resources from data level to knowledge level, so as to realize the orderly organization and knowledge utilization of resources. [Method/process] Using Stanford University's "Seven Steps" for reference, we innovated the ontology construction in oral history archives, and completed the theoretical model to practical operation through the steps of metadata standard selection, metadata extraction, ontology model design and ontology instantiation. [Result/conclusion] The limitation of the collection of oral history archives resources is broken through the knowledge organization mode of "ontology", and the related knowledge of people, event, time and place is reorganized, correlated and visualized, which promotes the deep development and value mining of resources.