Abstract: In recent years, the application of CPU-GPU heterogeneous system in the field of optical remote sensing image data processing has received wide attention. Firstly, the architecture and development of CPU-GPU heterogeneous system were introduced. Next, the process of optical remote sensing image data processing was introduced. Then, the application of CPU-GPU heterogeneous system in optical remote sensing image preprocessing, follow-up processing data processing was introduced. Finally, the application of CPU-GPU heterogeneous system in optical remote sensing image data processing system was analyzed and summarized. The analysis shows that the CPU-GPU heterogeneous system is feasible and has a wide prospect in the field of optical remote sensing image data processing, but still needs to solve the key problems such as parallelizing design and optimization of the algorithm, the load balance of CPU and GPU, which is of great significance to promote the application of the CPU-GPU heterogeneous system in the optical remote sensing image data processing.