Platelet size, measured as mean platelet volume (MPV), is associated with platelet reactivity. MPV has been identified as an independent risk factor for future stroke and myocardial infarction. The aim of this study was to determine the association of MPV with the development of stoke in patients with atrial fibrillation (AF). MPV, N-terminal pro B-type natriuretic peptide (NT-proBNP), and high-sensitivity C-reactive protein (hsCRP) were analysed in 200 patients with AF (mean age 69 years; 56% male). The primary endpoint was ischaemic stroke event. The mean MPV was 8.5 ± 1.0 fL and the median NT-proBNP was 1916.5 (IQR 810-4427) pg/mL. The median hsCRP was 0.47 (IQR 0.32-2.46) mg/dL. There were 14 stroke events during a mean of 15.1 months of follow up. Kaplan-Meier analysis revealed that the higher tertile MPV group (≥8.9 fL) had a significantly higher stroke rate compared to the lower tertile MPV group (<8.0 fL) (14.7% vs. 3.1%, log-rank: P = 0.01). A higher MPV was an independent predictor of stroke risk after adjusting for age, gender, and other CHADS(2) (congestive heart failure, hypertension, diabetes, and previous stroke or transient ischemic attack (TIA) history) score components (hazard ratio: 5.03, 95% CI 1.05-24.05, P = 0.043) in Cox proportional hazard analysis. When the MPV cut-off level was set to 8.85 fL using the receiver operating characteristic curve, the sensitivity was 71% and the specificity was 69% for differentiating between the group with stroke and the group without stroke. This value was more useful in patients with a low to intermediate traditional thromboembolic risk (CHADS(2) score <2). Furthermore, AF patients with an MPV over 8.85 fL had high stroke risk without anticoagulation, especially in the low thromboembolic risk group (Log-Rank <0.0001). The results of this study show that MPV was a predictive marker for stroke; its predictive power for stroke was independent of age, gender, and other CHADS(2) score components in patients with AF. These findings suggest that anticoagulation may be needed in patients with a high MPV, even if they have low to intermediate traditional thromboembolic risk (CHADS(2) score <2).