In recent years, China has successfully developed coal-based methanol to olefins (MTO) production processes and technologies, which has promoted the rapid development of the coal-to-olefins industry and guaranteed national energy security. The fluidized bed reactor is the core reaction device for the industrial production of methanol to olefins. It is of great significance to deeply understand the fluidization characteristics of the MTO fluidized bed through computational fluid dynamics. It can give more accurate guide for optimization and amplification of MTO fluidized bed. In this work, the multi-scale computational fluid dynamics (CFD) method which is coupling of bubble-based EMMS drag and traditional TFM is adopted to perform 3D simulation of the multi-phase flow behavior inside an industrial-scale MTO fluidized bed. This multi-scale CFD method involves in the influence of the bubble-based structure on the gas-solid drag coefficient. Thus, it can more accurately predict the "S-shaped" distribution of the axial particle concentration inside the MTO fluidized bed, which is consistent with experimental data. The predicted radical distribution of particle concentration presents the classic "core-annulus" flow structure. The predicted distribution of the averaged gas/particle axial-velocity in the radial direction is also mutually confirmed with the actual situation. This multi-scale CFD method significantly improves the predictive ability of the traditional TFM based on uniform drag for the macroscopic flow field. In future, the focus will be put on extending this multi-scale CFD method to the optimization design and reaction characteristics of MTO fluidized bed.