Scale-up of fluidized bed reactors has long been regarded as a big challenge in chemical reaction engineering. While traditional scaling theories are mostly based on hydrodynamics similarity, computational fluid dynamics (CFD) aided approach allows direct coupling between hydrodynamics and reaction factors and is expected to speed up the experiment-based scale-up process with lower cost. In this study, we aim to investigate the scale-up effects through simulations of a series of methanol-to-olefins (MTO) reactors of different sizes. The two-fluid model and energy-minimization multi-scale (EMMS)-based drag models, are combined in simulations. The fluidization characteristics in terms of flow structures, velocity distribution, mass fractions of gaseous product and coke distribution are presented against available experimental data for different-sized reactors. It is found that typical hydrodynamic features can be reasonably predicted, while the prediction of reaction behavior shows growing discrepancy with increasing reactor size. Possible reasons are discussed in the last section along with future work presented for scale-up studies. (C) 2017 Elsevier Ltd. All rights reserved.