Abstract: X-ray tomography technology(X-CT) can be used to observe the formation and decomposition processes of natural gas hydrate in porous medium, with the advantages of real-time, intuitive and lossless. However, the boundary region condition of different materials in sediment system is difficult to be determined due to the volume effect of particle edge image and low resolution, resulting in reducing the identification accuracy of gas hydrate occurrence in sediments. The Matlab platform image processing software was used for optimum processing of the CT images obtained from the experiment of methane hydrate in porous media. The results show that a more clear and distinguishable image can be obtained after optimizing the CT image with opening and closing operation, multi-value, gradient edge detection and image extraction methods, which can effectively help enhancing recognition for different boundary conditions of materials and improving the identification of methane hydrate occurrence in porous media.