Focusing on the problem of moving object detecting before tracking in image sequences with complex background, a new salient feature extraction algorithm that fuses the motion characteristics of objects is proposed. The algorithm fuses the gray-level and the detail and motion of the objects to form a robust salient image. The regions with salient multi-features are enhanced and the others are restrained, so moving objects will easily be detected in the salient image. The salience of the objects in the algorithm is derived from the local difference of multi-bottom features of the image, so the algorithm is robust. The experimental results show that the new algorithm is practical and stable, and has the strong anti-jamming capacity. The visual effect of the experiments demonstrates that this method can improve the signal-clutter ratio.