Abstract: Forest fog often leads to poor quality of collceted image information. In order to obtain clearer and complete images of forest defogging images and provide better data support and guarantee for forest monitoring, adaptive algorithm is used to preliminarily determine the forest fog content, light intensity and other environmental conditions. Then, the light and dark channel fusion algorithm is used to optimize the global atmospheric light, transimittance and other important parameters of the atmospheric model. Finally,Gaussian curvature filtering is used to deal with the halo effect caused by the light and dark channel prior algorithm. Light and dark channel fusion defogging algorithm based on adaptive image enhancement can effectively improve the quality of forest images, and obtain clear and complete forest defogging images with rich details and clear visual effects. The experiment evaluates the image effect scientifically from subjective standard and objective standard respetively, and the fusion improved algorithm obtained has good effect on forest image optimization.