Background: CHADS2 and CHA2 DS2-VASc scores are validated tools for assessing stroke risk in patients with atrial fibrillation. We aimed to investigate the correlation of CHA2DS2-VASc and CHADS2 scores with short-term prognosis in acute ischemic stroke patients with non-valvular atrial fibrillation (NVAF) in elderly patients. Methods: A total of 216 patients with non-valvular atrial fibrillation were enrolled in this study. Consecutive ischemic stroke patients with non-valvular atrial fibrillation who were hospitalized within 7 days after stroke were registered. Patients were divided into 3 groups on the basis of CHADS2 scores (0, 1, 2 to 6) and CHA2DS2-VASc score (0, 1, 2 to 9). And recovery was assessed by modified Rankin Scale (mRS) at 3 months follow-up (mRS <= 2 reflected good prognosis and mRS >= 3 implicated unfavorable outcome). After screening the risk factors affecting prognosis using univariate analysis, the correlation of CHADS2 and CHA2DS2-VASc scores for short-term prognosis was estimated using Logistic regression model. Results: A total of 206 cases (95.3%) were included in the study. The groups with CHADS2 scores of 0, 1, and 2-6 points contained 31, 83, and 92 cases, respectively, while the groups with CHA2DS2-VASc scores of 0, 1, and 2-9 points had 22, 78, and 106 cases, respectively. There were 89 (43.2%) patients with good prognoses and 117 (56.8%) with poor prognoses. The proportion of high-risk group was significantly higher (P<0.01) while that of low-risk group significantly lower as stratified by CHA2DS2-VASc score than by CHADS2 scores (P<0.01). High-risk (>= 2) CHADS2 and CHA2DS2-VASc scores are both associated with 3-month poor functional outcomes (CHADS2: OR 2.85, 1.32-4.05; CHA2DS2-VASc: OR 3.24, 1.32-6.98) after adjusting for baseline differences in age, sex and neurological impairment. The CHA2DS2-VASc score is better than the CHADS2 score in estimating 3-month stroke outcomes in acute ischemic stroke patients with nonvalvular atrial fibrillation. Conclusions: The CHA2DS2-VASc score appears to be a simple tool for identifying patients at lower risk of poor outcomes and serious cardiac complications within 3 months following ischemic stroke in patients with AF.