Shale gas is becoming an increasingly supplementary energy source because of its clean-burning and abundance. Economic gas production in shale requires the techniques of horizontal drilling and multistage hydraulic fracturing to create complex fracture network (CFN). How to accurately describe the characteristics of geometry and flow mechanisms of the CFN and select the most efficient approach for modeling are challenging. In this paper, a production forecasting model for multistage fractured horizontal well (MFHW) with CFN in shale is proposed based on the multiple interacting continua (MINC) theory (organic/inorganic matrix, natural fractures system) and lower-dimensional discrete fracture network (DFN) model (hydraulic fractures system). The model is designed to describe the unconventional flow mechanisms in shale system, such as fractal porous media and non-Darcy multiscale flow in ultra-tight matrix, ad-desorption on organic materials’ surface, rock un-consolidation within natural fractures, high-velocity turbulent flow near well range, and multiphase behaviors. We also propose a novel hybrid control volume finite element (CVFE) and finite-difference (FD) simulation method to obtain the numerical results of the model based on the unstructured 3D tri-prisms. The accuracy of the simulator is successfully validated and sensitivity analysis of some key factors (e.g.: fractal model permeability, Langmuir volume, heterogeneities of reservoir and fractures, well platform) are conducted to evaluate the impacts on production performance. Combing with the micro-seismic monitoring (MSM) data and engineering analyses, the DFN model is applied in Longmaxi shale formation to obtain the history matching with the field data and predict the production.