Spatial data infrastructure (SDI) is a framework of spatial data, metadata, users and tools that are interactively connected in order to use spatial data in an efficient and flexible way. As the development of network information technology and impact of global change, much attention is paid on SDI. However, SDIs have limited capability in dealing with ever-increasing amounts of high resolution data which can be distributed globally and heterogeneous extremely. One of the most important applications in grid is focusing on the collection of data resource. With the availability of large storage and computing facilities offered by Grid infrastructures, new opportunities are emerging. In the context of gird, the paper proposes the concept of “Data Grid Node” to shield the complexity process of heterogeneous spatial data source gridification, and introduces ways of automatic data grid entity building which can be used for the self-organization of distributed heterogeneous spatial data sources. The project “Data sharing infrastructure of global change”, funded by “National Research and Development Program of China (973 Program 2009CB723906)”, will target these issues.