Despite potential advantages of warfarin in reducing stroke risk, the percentage of patients with non-valvular atrial fibrillation (NVAF) using warfarin is relatively low. Our aim was to investigate the effects of stroke risk factors on warfarin, which was prescribed for inpatients with NVAF. A retrospective analysis was conducted using the data from 232, 237 and 251 NVAF inpatients obtained in 2003, 2008 and 2013 respectively from Beijing Anzhen Hospital, China. The Cochran-Armitage test was used to assess the trend of warfarin prescription in the NVAF inpatients. Logistic regression was used to analyze the association between stroke risk factors and prescribed warfarin. The results showed that inpatients with NVAF who were prescribed for warfarin lifted from in 2003 to in 2013. Although CHADS(2) score was not associated with the warfarin usage, coronary heart disease (CHD) and radiofrequency ablation (RFA) were negatively and positively respectively correlated with the warfarin use among NVAF inpatients. The study addressed anti-coagulation by using CHADS2 scoring to identify patient groups that are not receiving optimal medical therapy. Thought a substantial number of patients were under treated for thromboembolic prophylaxis in the management of NVAF, the work identified sub-groups such as those with CHD as being most at risk of under treatment.