The maritime search and rescue system which based on the helicopter plays an important role in the process of shipwreck. The current domestic maritime search and rescue flight team is not equipped with an infrared search and rescue system. It is imminent that the airborne infrared search and rescue system, which achieves fast and accurate positioning of the drowning. One of the difficulties is to achieve fast infrared image stitching. In this paper, on the basis of the existing algorithms, combined with the project requirements, a fast stitching algorithm for infrared images is proposed. Firstly, the original image is pre-treated to eliminate the effects of noise. Secondly, the image ROI area to be found by using of image similarity overlap and image matching is completed by using an adaptive algorithm based on the absolute difference. Finally, fusion of the continuous images is achieved by using the weighted average algorithms. The experimental results show that: the algorithm has a fast stitching speed and strong portability and practicality.