Abstract: Based on the similarity between the present image and background images,a method based on grey relational theory is proposed to detect the moving objects from video sequences．The video sequences are captured by the fixed cameras under various illumination conditions of indoor and outdoor scenes．By appropriately choosing the comparative sequences,the grey relational degree between them and the present frame is computed and moving targets are viably detected．The proposed method is not exigent to the background image,is insensitive to noise and to some extent is also robust to occlusion．