Abstract: Aiming at the pain points of railway scientific research enterprises, including scattered data storage, prominent data islands, poor data quality, and insufficiently released data value, this paper carried out adaptive optimization based on the Data Management Capability Maturity Model (DCMM) combined with the characteristics of railway scientific research businesses. The paper proposed a novel governance architecture covering data asset strategy, organizational management, system construction, process management, evaluation assessment, data resource utilization, data assetization, and governance tools. Meanwhile, the paper constructed a five-layer technical support system consisting of a data source layer, a data aggregation layer, a data base layer, a data governance layer, and a service application layer. It also analyzed the application value of the governance system in typical scenarios, including full-life cycle empowerment of scientific research, construction of high-quality training datasets, and panoramic business decision-making. The research results provide systematic theoretical references and implementable practical guidelines for the construction of data asset governance systems in railway scientific research enterprises, and help release the value of railway data elements and promote the high-quality development of the railway industry.