A large cohort study of high-risk population of stroke based on the real world is of great significance for stroke prevention and control. However, the data element structures, variable definitions and scopes of regional big data platforms are inconsistent, which will be an obstacle for data sharing, summary, and analysis among different regions. In this study, we formed an expert consensus on a unified minimum dataset standard for the cohort study of high-risk population of stroke, considering the categories and definitions of risk factors of stroke, and the existing database of the regional big data platforms. The consensus shall provide a reference for the comparison, integration, and sharing of real world data within and between regions, and play an important role in the cohort study on risk factors of stroke, as well as the implementation and evaluation of prevention and control measures.