A multifractal theory analysis of infrared image is proposed and the multifractal characteristics of edges in the infrared images are extracted. The singularity characteristics of each pixel in the image and its multifractal spectrum are calculated. Then pixels are classified by HVS(human visual system), which is more sensitive to the edge structure of image. From smoothness area to edge area, the pixels are weighted and enhanced. Infrared images are first analyzed using multifractal theory so that singularity α(x, y) of each pixel can be extracted from an infrared image. Multifractal spectrum f(α) is then estimated, which can reflect overall characteristic of an infrared image, and singularity exponent can reflect local characteristic of image so that edge and texture of an infrared image can be accurately extracted based on singularity of each pixel and multifractal spectrum. And finally the edge pixels are classified and enhanced in accordance with the sensitivity of human visual system to the edge profile of an infrared image. Experiments show that the image enhancement method is more available and highlights the human eyes' sensitive image area. The problem of low visibility and blurry edge is resolved efficiently and the enhancement image is more suitable to human's observing.