In patients with chronic heart failure (CHF), N-terminal pro–brain natriuretic peptide (NT-pro-BNP) predicted poor outcome. Clinical predictors of NT-pro-BNP and its usefulness in the presence of chronic kidney disease (CKD) are largely unknown. A total of 341 patients with stable CHF were enrolled, of whom 183 (54%) had CKD. During a follow-up of 620 ± 353 days, 57 patients (17%) experienced a cardiac event (cardiac death, need for extracorporeal assist device, or urgent cardiac transplantation), and 64 patients (20%) were rehospitalized because of worsening CHF. NT-pro-BNP was related to New York Heart Association functional class (R = 0.44, p <0.001) and inversely related to ejection fraction (R = −0.52, p <0.001) and glomerular filtration rate (R = −0.32, p <0.001). A cardiac event was independently predicted by NT-pro-BNP (hazard ratio [HR] 1.56, p <0.001), ejection fraction (HR 0.95, p = 0.018), and serum sodium (HR 0.89, p = 0.004). Using receiver-operator characteristic analysis, NT-pro-BNP ≥1,474 pg/ml best separated patients with or without cardiac events. In patients without CKD, outcome was significantly worse in patients with NT-pro-BNP >1,474 pg/ml in comparison to patients with NT-pro-BNP <1,474 pg/ml (event-free survival rate 0% vs 75%; p <0.001). In patients with CKD, outcome was also significantly worse in subjects with NT-pro-BNP >1,474 pg/ml in comparison to those with NT-pro-BNP <1,474 pg/ml (event-free survival rate 48% vs 93%; p <0.001). NT-pro-BNP independently predicted rehospitalization caused by worsening CHF (HR 1.26, p = 0.023), and a cut-off value of 1,474 pg/ml also separated patients with poor and intermediate prognosis in the CKD and non-CKD groups. In conclusion, NT-pro-BNP independently predicted morbidity and mortality in patients with CHF with and without CKD.