Introduction: We previously reported that a computational modeling-guided antiarrhythmic drug (AAD) test was feasible for assessing diverse AADs in patients with atrial fibrillation (AF). In this study, we took the virtual AAD test (V-AAD) in patients who took AADs after AF catheter ablation (AFCA). Hypothesis: Patients using an effective drug in the VAAD test will have fewer AF recurrences than patients using an ineffective drug. Methods: This single-center retrospective study included 246 patients (72.8% male, 60.7±10.2 years of age, 38.6% paroxysmal AF) prescribed AADs within 3 months after AFCA. Using realistic computational modeling, we evaluated the effects of five AADs (amiodarone, sotalol, dronedarone, flecainide, and propafenone; 2 doses for each drug). Clinical AADs (C-AAD) were chosen at the discretion of the physicians blinded to the V-AAD test. We defined the effective V-AAD as the V-AAD that terminated AF or converted AF to atrial tachycardia (AT) and the best V-AAD as the V-AAD that ended virtual AF the fastest. We compared AF recurrence rates after the AAD prescription depending on the results of the V-AAD test. Results: The ineffective and best V-AAD were administered in 64 (26.0 %) and 70 patients (28.5%). The recurrence rate within a year after using the ineffective, effective, and best V-AAD were 46.9%, 35.7%, and 25.7%, respectively (log-rank p=0.013). The recurrence rate within a year in patients with at least one V-AAD terminating AF and those without was 33.9% and 50.0%, respectively (log-rank p=0.034). The use of best V-AAD (OR 0.43, 95% CI [0.18-0.96]; p=0.042, vs. ineffective V-AAD) is an independent predictor of AF recurrence within a year of using AAD after AFCA. Conclusions: The digital twins-guided V-AAD test was feasible for evaluating the efficacy of multiple AADs in patients with AF who had a high chance of recurrence after AFCA. We need a prospective randomization study to assess the prediction power of the V-AAD test.