Abstract: Fractures in coal are the main channels for the migration of coalbed methane and are the direct factor that affects and controls the permeability of coal reservoirs. Tectonically deformed coals are generally developed in China, and the characteristics of fractures in different types of tectonically deformed coals are significantly different, and they have different effects on the permeability of coal reservoirs. Therefore, it is necessary to study the fracture and permeability in the tectonically deformed coal. In this study, the middle and high-rank tectonically deformed coal samples were selected from the Pingdingshan 13th Mine and the Jincheng Hudi Mine. Firstly, the columnar coal samples were made by epoxy resin to solve the problem of difficulty in the sample preparation of tectonically deformed coal.Secondly, by means of scanning electron microscopy, micro-CT scanning and digital image processing technology, the differences of development, distribution, morphology and structure of aperture greater than 15 μm in the two coals were analyzed. Finally, a permeability prediction model was established based on the Poiseuille's law, triaxial permeability experiment and fracture structure parameters of CT scanning. The mechanism of fracture structure controlling permeability was discussed. The results show that large fractures with simple morphology are mostly developed in the weak tectonically deformed coal(tend to primary structural coal) under the scale of aperture greater than 15 μm, and the fracture connectivity and connection strength are poor, while the micro-fractures in the strong tectonically deformed coal are complex in morphology, and the fracture connectivity and connection strength are good. The fracture porosity of coal with the strong tectonic deformation is larger than that of coal with weak tectonic deformation. Permeability decreases in a negative exponential trend with the increase of effective stress. The permeability of coal reservoirs is controlled by multiple factors, including fracture aperture, number, volume, connectivity and connection strength, of which aperture has the most significant influence. In addition, considering the factors of controlling permeability and the fractal dimension and roughness of fractures, a tubular fracture permeability model was established, and the accuracy and applicability of the model in predicting permeability at this scale were verified.