Background: Previous data are controversial about the association of renal artery stenosis (RAS) with clinical outcome in patients with heart failure. Definition of RAS in previous studies might not be appropriate. By definition of RAS with renal duplex sonography, we investigated the association of RAS with clinical outcome in patients with heart failure. Methods: In this retrospective study, we identified 164 patients with heart failure (New York Heart Association classification >= II; left ventricular ejection fraction < 50%) who had received renal duplex sonography during hospital stay. RAS was defined as renal-aortic ratio >= 3.5 or a peak systolic velocity >= 200 cm/s (or both), or occlusion of the renal artery. Categorical data of patients were compared using the Chi-square test or Fishers exact test. Cox proportional hazards regression modeling technique was used to investigate the prognostic significance of possible predictors. Results: Finally, 143 patients were enrolled. Median follow-up time was 32 months (1-53 months). Twenty-two patients were diagnosed as RAS by renal duplex sonography, including 13 unilateral RAS (3 left RAS, 10 right RAS) and 9 bilateral RAS. There were more all-cause mortality and cardiovascular death in patients with RAS than patients without RAS. By multivariate analysis, RAS was a significant predictor for all-cause death and cardiovascular death (hazard ratio [HR] = 4.155, 95% confidence interval [CI]: 1.546-11.164, P = 0.005; and HR = 3.483, 95% CI: 1.200-10.104, P = 0.022, respectively). As for composite endpoint events, including death, nonfatal myocardial infarction, ischemic stroke or intracranial hemorrhage, rehospitalization for cardiac failure, and renal replacement therapy, only angiotensin-converting enzyme inhibitor or angiotensin-receptor blocker was significant predictor. RAS was not a significant predictor for composite endpoint events. Conclusions: Our data suggested that RAS is associated with a poorer clinical outcome in patients with heart failure.