Abstract: [Purpose/significance] At present, the fine- grained humanistic knowledge (emotions, concepts, thinking,styles, etc.) in the field of intangible cultural heritage is widely scattered in unstructured texts of multiple sources, and has not been effectively organized. How to design a model for automatic transformation from large- -scale intangible heriage texts to humanistic knowledge is an important guide for the construction of humanistic knowledge graphand its application.[Method/process1 In this paper, we explore the domain know ledge organization model for unstructured ICH texts, focusing on the technical characteristics of knowledge graphs in semantic association parsing of texts and semantic data linking, introducing natural language processing techniques for semantic association parsing of humanistic knowledge in ICH texts, and promoting semantic data linking and knowledge services based on semantic fusion.A technical path is demonstrated by taking“Guqin Art”in the ICH list of the United Nations as an example, ineluding: metadata feature analysis and humanistic ontology modeling, semantic association parsing of text incorporating linguistic features of Chinese characters under“cold start", multi- source knowledge link under semantic fusion, and deduction and knowledge discovery basing knowledge ontology. [Result/conclusion]We aim at constructing a complete set of knowledge organization model and automated realization path from unstructured text to structured knowledge and open shared knowledge base. The effective use of semantic technology refines the granularity of domain knowledge, and the semantie description and association of knowledge enhance the scope of data linking, thus laying the foundation for deeper ICH humanistic knowledge services.avoidance.