To evaluate the cost-effectiveness of rivaroxaban vs. vitamin K antagonists (VKAs) such as warfarin, for stroke prevention in patients with non-valvular atrial fibrillation (NVAF) in France, using real-world (RW) evidence, from a national health insurance perspective. A Markov model considering a 30-year time span with cycle lengths set to 3 months was developed. Patient data for characteristics of the model population were obtained from a RWE study, considered as a relevant source of characteristics for French patients with NVAF. The model used RW baseline clinical event data drawn from different sources to inform expected event and persistence rates for VKA and rivaroxaban treatment arms. A meta-analysis of RW studies provided treatment effect data for rivaroxaban compared to VKA. Model outcomes included: costs (drug costs, clinical event costs, and VKA monitoring costs), quality adjusted life-years (QALY); life-years (LY) gained; incremental cost per QALY; incremental cost per LY; number of stroke events; number of gastro-intestinal bleeds; and number of myocardial infarctions. Sensitivity analyses were run to better understand the results drivers and robustness. In the base-case analysis, the incremental total cost was €826 and total incremental QALYs and LYs were 0.103 and 0.114, respectively. The resulting incremental cost/QALY and incremental cost/LY were €8,015 and €7,246, respectively. The results were most sensible to inclusion of treatment-specific disutility, and RW baseline clinical event data. Although there is no official cost-effectiveness threshold in France, this study demonstrates that rivaroxaban can be considered cost-effective vs VKA in a real-world setting, using a French national insurance perspective.