Dense reactive flows have been extensively applied in industrials, such as coal/biomass gasification and combustion in fluidized beds. To achieve an optimized performance for the design and operation of the system, the micro- and macro- behaviors of dense reactive flows including solid mixing, cluster, regime transition etc. should be well-understood. However, deep insights into these facilities challenge the capacity of measuring techniques due to the harsh conditions within the apparatus. Although efforts have been taken to investigate reactive fluidized beds under various conditions, there is still a lack of thorough understanding on the interplays among the gas-solid hydrodynamics, heat transfer, and chemical reactions. It is noted that most previous numerical studies on the reactive fluidized beds were based on the continuum-based simulations, in which it is difficult to achieve the particle-scale information. Comparatively, the coupling of computational fluid dynamics (CFD) and discrete element method (DEM) in the Eulerian-Lagrangian framework has emerged as a powerful tool to provide predictions with detailed particle-scale information, which facilitates the investigations of underlying mechanisms in fluidized beds. This talk will present an overview of the work in this field at Zhejiang University. It covers particle-resolved direct numerical simulation (PR-DNS) and combined computational fluid dynamics (CFD) and discrete element model (DEM) as well as coarse-grained models on dense reactive flows in fluidized beds from the perspective of sub-particle, particle, and particle parcel scale. Heat mass transfer and chemical reactive have been incorporated into the framework. These methods have been validated against experimental data at different scale levels, and applied to investigate the multi-physics coupling characteristics in dense reactive flows.