Taking the unique advantage of large-scale ReaxFF MD simulation in unravelling varied reaction pathways within one simulation, the process of methanol-to-olefin (MTO) was investigated by the combination of high-performance computing and cheminformatics-based reaction analysis. The initial stage model and the indirect pathway model for MTO were constructed to get a more complete scenario of the whole catalyzed process. The obtained dynamic profiles of dominant products, intermediates, and detailed reactions indicate that the MTO processes catalyzed over ZSM-5 with different acidic properties exhibit similar proceeding stages but with significant different pathways. The MTO process catalyzed by hydrated ZSM-5 with Al enhances the methane-formaldehyde route, H and CH3 transfer reactions, and oxidation reactions of surface methoxy species during the direct pathway, leading to CO2 production and CH2O consumption, which is barely taking place at the model catalyzed by zeolite without Al. Semiquantitative reaction networks with conversion ratios for C2H4 and C3H6 generation revealed that the MTO process is a product of multiple pathways with dozens of intermediates. The rich information, especially the comprehensive reaction network obtained from ReaxFF MD simulations would complement the experimental results and those of other theoretical research and be helpful for further development of highly efficient MTO processes.