Aiming at the problem of low detection accuracy and poor noise robustness of moving target detection algorithm, this paper studies a moving target detection system combining three-frame difference method and Gaussian model, which adopts inter-frame difference and threshold segmentation to extract the moving region in the image, and further enhances the accuracy and practicability of the three-frame difference method. Experimental results show that compared with the single TFDM algorithm, this system has higher detection accuracy and better robustness to noise than the traditional inter-frame difference method and template matching method.