Abstract: Due to the numerous production links, complex organizational structures and processes in nuclear power, a unified industrial big data platform at the nuclear power group level should be established to eliminate data silos, integrate massive data resources, and improve data utilization efficiency. In this study, a multi-source heterogeneous data access framework adapted to the nuclear power industry protocol is adopted to aggregate the cross-regional nuclear power industry data. According to the characteristics of nuclear power industry data, the coding and storage standards of nuclear power industry data are defined, and the efficient storage of massive data is realized. Based on metadata management, data extraction, data processing, data storage, data scheduling, quality audit, data service and other links are coordinated, and an end-to-end and closed-loop data governance and control mechanism is studied. This method was used to govern nuclear power data and open up assets, achieving full lifecycle management of nuclear power data. Represented by China Nuclear Power Data Center and six major nuclear power bases, the construction and engineering practice of big data platforms have formed a data sharing and collaboration channel between the group's data center and the edge side; We have gathered and processed data on China's nuclear power production and management, established data standards, and achieved effective integration and processing of data. Therefore, the nuclear power industry big data platform and data processing methods established in this study can provide support for the digitalization of nuclear power, have broad application prospects, and provide technical references for building a nuclear power data ecosystem.