Abstract: A new method for real-time color video fusion based on adaptive registration parameter update was proposed. First, the registration parameters of each frame image were updated adaptively and each visible image and infrared image was registered. Secondly, linear gray level fusion method was used for the registration image, and the appropriate color target image was selected and the color transfer of color target image to initial gray fusion image in YUV color space was completed. Finally, the color fusion image was gotten. Experimental result shows that the color-transferred image has better natural color perception than the conventional false color image. Moreover, the algorithm is fast and convenient for real-time digital video processing.