For the problem of infrared dim-small target tracking,a fast matching tracking method is put forward based on nonlinear quantization probability density model.The characteristics of infrared dim-small targets related to the gray characteristics and multi-frame motion are first analyzed.α-β filter is then introduced to track motion infrared dim-small targets and the centroid of the target in the next frame image is predicted.Meanwhile,a nonlinear quantization probability density model for infrared dim-small targets is proposed,and the real target is tracked by template matching in the predicted neighborhood.Experimental results demonstrate the proposed method for infrared dim-small target tracking has a great advantage over the present methods in accuracy and speed,so it is especially valuable for engineering application.