For infrared image, how to enhance the detail and suppress the noise while compressing the dynamic range is an important issue. An improved infrared image adaptive enhancement method was proposed. Firstly, a parameter adaptive guided image filtering method was designed, based on the guided image filtering, the original infrared image was divided into the basic layer and the detail layer; based on the gray distribution, a new histogram projection method with adaptive threshold was designed to compress the dynamic range and enhance the contrast of the basic layer; then the linear coefficient of the adaptive guided filter was used to enhance the detail layer and suppress the noise; finally, the enhanced infrared image was obtained by adaptive fusion of the enhanced basic layer and detail layer. The experimental results show that compared with several good algorithms, such as the contrast limited adaptive histogram equalization and the detail enhancement for high-dynamic-range infrared images based on guided image filter, the proposed algorithm processed image detail is richer, noise suppression effect is stronger, better visual effect, and the algorithm is more flexible, without adjusting parameters to deal with a variety of scenarios.