This study aimed to evaluate how large the variation between measurements of N-terminal pro-B-type natriuretic peptide (NT–pro-BNP) can be in patients with clinically stable heart failure (HF). The fluctuation of NT–pro-BNP in patients with HF has not been previously evaluated. Patients with chronic, stable HF were evaluated at 3-week intervals. The relative (percentage of baseline) difference between the 2 NT–pro-BNP measurements and their mean were calculated. In total, 118 double measurements of NT–pro-BNP were obtained. Changes in NT–pro-BNP measurements were determined according to quartiles of mean NT–pro-BNP. The median baseline NT–pro-BNP level was 1,323 pg/ml (25th and 75th percentiles 433 and 2,258), and the median second NT–pro-BNP level was 1,394 pg/ml (25th and 75th percentiles 475 and 2,353) (p = 0.34). Among pairs of measurements in the first quartile (NT–pro-BNP <450 pg/ml), the 5th and 95th percentiles of the distribution of relative differences were −86% and +197%. In the second (450 to 1,300 pg/ml), third (1,300 to 2,300 pg/ml), and fourth (>2,300 pg/ml) quartiles, the 5th and 95th percentiles of relative differences were −42% and +78%, −29% and +38%, and −30% and +37%, respectively, of the baseline values. In conclusion, the results support the clinical use of NT–pro-BNP in the monitoring of patients with HF with NT–pro-BNP levels >1,300 pg/ml. In these patients, variations between 1/3 less and about 40% more can be expected without clinical improvement or deterioration.