Background and Purpose— Elevated serum levels of brain natriuretic peptide (BNP) have been associated with cardioembolic stroke and increased poststroke mortality. We sought to determine whether BNP levels were associated with functional outcome after ischemic stroke. Methods— We measured BNP in consecutive patients aged ≥18 years admitted to our stroke unit between 2002 to 2005. BNP quintiles were used for analysis. Stroke subtypes were assigned using Trial of ORG 10172 in Acute Stroke Treatment criteria. Outcomes were measured as 6-month modified Rankin Scale score (“good outcome”=0–2 versus “poor”) as well as mortality. Multivariate logistic regression was used to assess association between the quintiles of BNP and outcomes. Predictive performance of BNP as compared with clinical model alone was assessed by comparing receiver operating characteristic curves. Results— Of 569 patients with ischemic stroke, 46% were female; mean age was 67.9±15 years. In age- and gender-adjusted analysis, elevated BNP was associated with lower ejection fraction ( P <0.0001) and left atrial dilatation ( P <0.001). In multivariate analysis, elevated BNP decreased the odds of good functional outcome (OR, 0.64; 95% CI, 0.41–0.98) and increased the odds of death (OR, 1.75; 95% CI, 1.36–2.24) in these patients. Addition of BNP to multivariate models increased their predictive performance for functional outcome ( P =0.013) and mortality ( P <0.03) after cardioembolic stroke. Conclusions— Serum BNP levels are strongly associated with cardioembolic stroke and functional outcome at 6 months after ischemic stroke. Inclusion of BNP improved prediction of mortality in patients with cardioembolic stroke.