MODIS data is the world’s best remote sensing data of free receiving. It plays an important role in agriculture science and application. However, there is a serious problem about duplicated establishment of MODIS data receive station in China, and it has brought a tremendous waste of money. This paper presents research work on utilizing data grid technology to integrate heterogeneous distributed MODIS data sources. We studied and implemented a three-layer structure for data interoperation, and designed the extensible XML-based spatial data access language to support cross-border data computation. We developed the MODIS data catalogue service to facilitate users and improve system performance. The research helps to release the conflict between duplicated establishment of MODIS receiving station and less effectively data sharing.