Abstract: In order to give an efficient evaluation of the quality of video images, based on the analysis of the existing algorithms of evaluation, this article proposes an realization of these algorithms under the Stream Processor Parallelenvironment and program with the technology of CUDA. The result shows that the speed of parallel computing in the NV Geforce GTX280 is about 30-time of the serial algorithm running in the CPU. This proves that CUDA technology has great potential in the digital image processing and particularly the CPU-intensive mission processing.