Abstract: In order to realize selective encryption of infrared images, an infrared image selection algorithm based on multi-feature difference detection and joint control mapping is proposed. Initial infrared image is enhanced by introducing segmented Sine transform to highlight true infrared target. Then target decision factor is designed by considering gray difference between target and background, and infrared background suppression mechanism is constructed by segmenting structure elements of Top-Hat transform to filter clutter and noise. Multi-feature difference detection model is established by using gray level, contrast and similarity degree to extract region of interest of real target and suspicious object. Joint control chaotic mapping is designed by Logistic map, Tent map and chebysbev map to permutation region of interest. Finally, gravity model is introduced to spread pixels in region of interest for complete infrared target selection encryption. Experimental results show that compared with existing local encryption mechanism, entropy of cipher text information is 7.982 6, which can be used for local selective encryption of infrared images.