Aim The population included in randomized controlled clinical trial of ROCKET AF and observational studies were different, and the effectiveness and safety of rivaroxaban in stroke prevention of atrial fibrillation (AF) varied among studies. The aim of this study was to estimate the real-world outcomes of rivaroxaban in AF patients in a relatively accurate way. Methods A discrete event simulation (DES) model was proposed to predict the counterfactual outcomes of ROCKET AF that would have it been conducted in broader observational study populations. The hypothetical cohorts of patients were generated using Monte Carlo simulation. The DES model structure was built based on disease progression and possible outcomes of AF. Cardiovascular events were recorded during the simulated two-year follow-up period. Results The results showed lower predicted rates of stroke/ systemic embolism (SE) and major bleeding in three observational studies than those in simulated ROCKET AF. The simulated stroke/SE incidence was 1.097-1.318 per 100 patient-years and simulated major bleeding incidence was 2.804-3.238 per 100 patient-years in observational studies. The risk difference of stroke/SE and major bleeding was similar among predicted outcomes of the three observational studies. Most simulated hazard ratios (HRs) were close to the corresponding observed HRs. Conclusion The simulated incidence of stroke/SE incidence and major bleeding might reflect the real-world event rate in AF patients. Even some differences existed in the absolute rates of stroke/SE and major bleeding between observed and simulated studies, the results confirmed similar effectiveness and safety to ROCKET AF comparing rivaroxaban and warfarin in AF patients.