In order to overcome the disadvantage of the background difference method and the frame difference method and improve the accuracy, realtime and detection efficiency for the detection of moving target, a method was presented that combines the continuous frame difference method with the background difference method. The proposed method firstly used continuous frame difference method to obtain continuous difference images. Then a linear adaptive filter or an nonlinear median filter was adopted to obtain the background image and get them subtraction. Finally a threshold segmentation technology was used to enhance the detected moving target. Experiment results demonstrate that the proposed algorithm can effectively avoids the phenomenon of the “missed” and “false alarms”, etc., and the detection efficiency and accuracy are improved greatly.