Abstract: A method of dim infrared target detection in complex background based on adaptive lateral inhibition network is presented, in order to improve the efficiency of long-range dim infrared target detection. The lateral inhibition network math model is established, and anisotropic Gaussian filtering is used to determine the coefficient of the lateral inhibition network. The scale of the filter’s u-axes is determined by contrast-dependent-spread filter model, while the scale of the filter’s v-axes is determined by intensity-dependent-spread filter model. Compared with traditional test methods, the method can effectively suppress the background, while enhancing the target.