Abstract: The paper studies the data category and data domain boundary from the perspective of concept domain, clarifies its multi-dimensional types and attributes, and helps to classify and standardize data management, as well as clarifying the data guarantee strategies of different attributes, and specifically implementing the Measures for Scientific Data Management. Some related concepts, definitions and expressions about data and scientific data are extracted from relevant literature, sorted and interpreted according to their production process, release form, function, discipline attribute, responsibility and equity ownership, so as to form the data type and dataset attribute boundary from the perspective of data domain concept. According to the relevant provisions in the Measures for Scientific Data Management, specific data domain boundaries and management rules of different categories and attributes are put forward to provide detailed strategies for the full implementation of the Measures for Scientific Data Management. In particular, the “boundary” definition in natural science data and social science data is conducive to the collection, management and implementation of scientific research data, and also to the rational reuse and sharing of the ownership of data fields with different attributes.