The institution is an important descriptive element of the literature. Identifying and tracking these institutions present challenges,as they often merge and split,acquire or are acquired by other institutions,change their names and locations. Institutions may be known by a variety of names over time even within one language and their hierarchical structures are also complex. So it is difficult to analyze institution knowledge unit in information retrieval,knowledge navigation,scientific research management and evaluation etc. In order to ensure the uniqueness of the institution,domestic and foreign organizations and scholars have carried out the construction of institutional authority file and explored institutional name disambiguation and normalization algorithm.The paper tries to construct knowledge organization system centered on institutional authority file to ensure the uniqueness of institutional name,the orderliness of the institutional inter hierarchical relationship,and the standardization of the relationship between institutions. The institutional authority file not only includes the first level organization but also involves fine grained organizations,such as the second level and third level sub units,and then realize the standardized control and organization of all levels of institutions and related knowledge resources.The paper focuses on the overall construction architecture,construction framework,construction ideas,key problem and technologies of institutional authority file. It analyses usage requirements and provides several possible future service or application models and update and maintenance strategies. We hope knowledge organization system centered on institutional authority file can provide support for the comprehensive description and disclosure of institutional attributes and relationships in the big data environment. 6 figs. 1 tab. 23 refs.