[Purpose/significance] The rapid development and wide application of information technology certainly promote the reform of Dunhuang studies, which make the utilization of Dunhuang studies and its knowledge resources more convenient. Because of the traditional knowledge platform and retrieval method of databases, the value mining of knowlegde of Dunhuang documents images is not complete. Therefore, this paper combs the research of Dunhuang documents images and gives a full semantic description for the depth and breadth of their knowledge, exploring the multiple values of Dunhuang documents images. [Method/process] Through the consideration and semantic description of present researches of Dunhuang documents images, hierarchical model of semantic features was designed in light of their semantic features. The organizational model was constructed by combing intelligent+ and linked data. And this paper deeply analyzes knowledge association of the Dunhuang documents images and the data collection layer, the semantic description layer, the data association layer and the resource application layer of the hierarchical model. [Result/conclusion] The paper puts forward the practical value of the knowledge association model of Dunhuang documents, which can open the connection channel with outside data, improve the retrieval intelligentized quality of Dunhuang documents images, and enhance its depth and breadth of the knowledge service.