Shale gas has become one of the primary energy resources during the past few years, and its impact has been profound in many countries. Hydraulic fracturing treatments are required for the development of shale gas reservoirs, and the consequent hydraulic fractures usually connect with the original small-scale natural fractures forming complex fracture networks in these reservoirs. Therefore, a model for numerical simulation, which is capable of accurately modeling naturally and hydraulically fractured reservoirs, is essential in optimization and management of such reservoirs. In this paper, we develop a comprehensive model that couples embedded discrete fractures, multiple interacting continua, and geomechanics to accurately simulate the fluid flow in shale gas reservoirs with multiscale fractures. Large-scale hydraulic fractures are described by an embedded discrete fracture method, while middle-scale and small-scale natural fractures are modeled by a multiple interacting continua method. Usually, the connection of matrix natural fractures hydraulic fractures wells is considered as the main pathway for the gas flow from a reservoir to a production well. However, geomechanics effects are significant in fractured reservoirs, which may lead to the closure of fractures and a dramatic decrease in gas conductivity in the fractures during the depletion of pressure. When the permeability of fractures is close to that of matrix, the gas production directly from the pathway of matrix hydraulic fractures wells as well as the fluid flow in shale matrix cannot be ignored. To accurately predict the fluid flow and well performance, the geomechanics effects, and all the possible connections between different regimes must be taken into account. We implement this comprehensive model in our in-house reservoir simulator and study its behavior by numerical experiments. According to the numerical simulation results, accurate and comprehensive production prediction is performed, and reasonable physical phenomena is captured. Sensitivity studies are also-performed to show the impacts of different parameters on the prediction of well performance.