Image enhancement can be divided into two kinds: global enhancement and local enhancement. Current image enhancement techniques based on local enhancement cannot accurately segment the target area and background, and it is difficult to enhance the segmentation region adaptively. In this paper, a region-adaptive multi-scale strong light fusion algorithm is proposed for infrared image enhancement. Firstly, semantic segmentation technology is used to divide the target area and background area. Then, the improved multi-scale strong light fusion algorithm is used to enhance each area adaptively. The experimental results show that the enhancement effect of the proposed algorithm is better than that of the current conventional algorithms, and the visual effect of image enhancement is more realistic.