Abstract: The existing heterogeneous clutter simulation methods based on geographic information usually use grid mapping method to partition geographic information, and then traverse the scattering units within each grid to calculate the scattering coefficient of the grid. This type of method is not conducive to the unified use of multi-source geographic information, and the division of grids and traversal of scattering units reduce the real-time performance of the simulation. Addressing the above issues, a fast simulation method facing simulation application for heterogeneous clutter, which is based on resampling and preprocessing multi-source geographic information according to distance-azimuth, is proposed in this paper. The proposed method avoids the need for grid cell division and traversal of scattering bodies within grid cells during the real-time computation phase. To achieve that, a fast heterogeneous clutter simulation framework based on graphics processing unit for space-based radar is designed. The experiment results show that the proposed method effectively improves the speed of heterogeneous clutter simulation under high simulation model accuracy conditions, and has good engineering application value.