Since OpenAI released the ChatGPT large model, all industries have rushed to apply large models in their practical work. However, as industries have used large models more deeply, they have found that these general-purpose large models have difficulty solving the application problems of real work. Therefore, this project takes archival data governance as its research object and uses the strong data processing capability of large models to carry out research on archival data governance based on a vertical-domain large model for archives. First, it analyzes the current status and problems of archival data governance.. Second, based on the strong data processing capability of the vertical-domain large model for archives, it carries out research on archival data quality optimization.. Third, it studies how the vertical-domain large model for archives can enhance the value of archival data, enhancing that value at the knowledge level.. Finally, it builds a data governance platform that integrates archival data quality optimization and value enhancement functions, and establishes a security mechanism to ensure data security. The research results will provide an exemplary reference for data governance in the archives sector.