Abstract: The deep coal seam has high free gas content and abundant resources, which is an important potential resource for natural gas storage and production. The gas content of deep coal seams, particularly the free gas content, is significantly influenced by the sealing condition of roof. Based on the micro-resistivity imaging logging technology, the roof lithology identification of certain coalbed methane production area in the eastern margin of Ordos Basin is carried out, the identification chart of coal seam roof lithology and fracture is established, the direction of ground stress is analyzed, and the favorable roof combination type is put forward. The result show that combining conventional logging data and electrical imaging characteristics, 10 types of lithology such as conglomerate, sandstone and mudstone are effectively distinguished based on static electrical image, corresponding to different Gamma and acoustic time logging results. The coal seam roof in the study area is dominated by mudstone, followed by sandstone. The cracks in the roof of the coal seam mainly develop high-conductivity joints and induced joints. The high-conductivity joints are shown as sine curves, including open joints and muddy filling joints. The induced joints are arranged in feathers or geese, which can be used to determine the direction of stress. Combined with the thickness of roof lithology and the development of cracks, three types of coal seam roof combinations with different sealing properties can be divided. Among them, the electrical imaging image of type I roof is generally bright and the cracks are less developed. Type II roof develops more cracks. Type III roof image shows that the cracks are cut and the image is not clear. Based on the above work, the distribution of three types of roof combination conditions in the study area is divided. In the deep area, it is mainly I and II, which has good sealing performance. The above work clarifies the display characteristics of different lithologies in electrical imaging logging, and proposes a deep coal seam roof sealing evaluation method based on micro-resistivity imaging logging. The related work can effectively support the optimization of deep coalbed methane dessert zones and well location deployment.