The methanol to olefins (MTO) process has been successfully commercialized in China and will potentially become an important route for light olefins production. In this work, a modeling approach is presented for MTO fluidized bed reactor design and operation optimization. The two-fluid model (TFM) where the solid phase shear viscosity and solid phase pressure are derived from kinetic theory of granular flow has been used to model the solid gas two-phase flows. The interphase drag force is calculated by either the traditional Gidaspow model or a recently developed EMMS-bubble model. The simulation study has been performed for a fluidized bed reactor in a 16 kt/a DMTO unit. It has been shown that the Gidaspow model cannot predict a stable dense bed, while the EMMS-bubbling model could simulate the solid fraction distribution in the reactor reasonable well. A reaction model based on the simple MTO reaction kinetics has been implemented to test the effectiveness of the model approach. The simulation results show that the methanol is converted rapidly just above the gas inlet. The selectivity of ethylene and propylene however are underpredicted, while the selectivity of CO2 and other products are overestimated. It is suggested that a further extension of the EMMS model to a turbulent fluidized bed is important in order to get more quantitative results. Also the MTO reaction kinetics for a commercial DMTO catalyst, in which the coke formation kinetics should be included, is highly desired.