Background: To investigate the efficacy and safety of hirudin plus aspirin therapy compared with warfarin in the secondary prevention of cardioembolic stroke due to nonvalvular atrial fibrillation (NVAF). Methods: Patients with cardioembolic stroke due to NVAF were prospectively enrolled from 18 collaborating hospitals from Dec 2011 to June 2015. Fourteen days after stroke onset, eligible patients were assigned to the hirudin plus aspirin group (natural hirudin prescribed as the traditional Chinese medicine Maixuekang capsule, 0.75 g, three times daily, combined with aspirin 100 mg, once daily) or the warfarin group (dose-adjusted warfarin targeting international normalized ratio (INR) 2-3, with an initial daily dose of 1.25 mg). Patients were followed up at 1, 2, 3, 6, 9, and 12 months after stroke onset. Time in therapeutic range (TTR) was calculated according to Rosendaal methodology to evaluate the quality of INR management in the warfarin group. The primary efficacy endpoint was the recurrence of stroke within 12 months after stroke onset. Safety was assessed as the occurrence of the composite event "intracranial hemorrhage and other bleeding events, death, and other serious adverse events". The Cox proportional hazard model and Kaplan-Meier curve were used to analyze the efficacy and safety events. Results: A total of 221 patients entered final analysis with 112 patients in the hirudin plus aspirin group and 109 in the warfarin group. Over the whole duration of our study, TTR for patients taking warfarin was 66.5 % ± 21.5%. A significant difference was not observed in the recurrence of stroke between the two groups (3.57% vs. 2.75%; P = 0.728). The occurrence of safety events was significantly lower in the hirudin plus aspirin group (2.68% vs.10.09%; P = 0.024). The risk for efficacy event was similar between the two groups (hazard ratio (HR), 1.30; 95% confidence interval (CI), 0.29-5.80). The safety risk was significantly lower in the hirudin plus aspirin group (HR, 0.27; 95% CI, 0.07-0.95). Kaplan-Meier analysis revealed significant difference in the temporal distribution in safety events (P = 0.023) but not in stroke recurrence (P = 0.726). Conclusion: Significant difference in efficacy was not detected between warfarin group and hirudin plus aspirin group. Compared with warfarin, hirudin plus aspirin therapy had lower safety risk in the secondary prevention of cardioembolic stroke due to NVAF.