Infrared small target detection has been the hot and difficult research of image processing, influenced by many factors, infrared small target images have the problems of low signal to noise ratio (SNR), poor-contrast and easy to be covered by background. Infrared image characteristics in sky background are researched. According to the characteristics of image scene, morphological algorithm is used to preprocess the original image and suppress noise influence. And then, using the correlation between pixels, if the same target cell, gray variance is generally gentler in horizontal and vertical directions, and the diagonal neighboring pixel difference information is used to perform infrared small target detection to noise suppressed images. Simulation results show that the algorithm is simple and convenient and can effectively extract targets.