Although exercise oscillatory ventilation has emerged as a potent independent risk factor for adverse prognosis in heart failure, it is not well known whether cardiac rehabilitation can improve oscillatory ventilation. In this study, we investigated the magnitude of oscillations in ventilation before and after cardiac rehabilitation in chronic heart failure patients with exercise oscillatory ventilation. Cardiac rehabilitation (5-month program) was performed in 26 patients with chronic heart failure who showed an oscillatory ventilation pattern during cardiopulmonary exercise testing (CPX). After the 5-month rehabilitation program was completed, the patients again underwent CPX. To determine the magnitude of oscillations in ventilation, the amplitude and cycle length of the oscillations were calculated and compared with several other parameters, including biomarkers that have established prognostic value in heart failure. At baseline before cardiac rehabilitation, both oscillation amplitude (R = 0.625, P < 0.01) and cycle length (R = 0.469, P < 0.05) were positively correlated with the slope of minute ventilation vs. carbon dioxide production. Plasma BNP levels were positively correlated with amplitude (R = 0.615, P < 0.01) but not cycle length (R = 0.371). Cardiac rehabilitation decreased oscillation amplitude (P < 0.01) but failed to change cycle length. The change in amplitude was positively correlated with the change in BNP levels (R = 0.760, P < 0.01). Multiple regression analysis showed that only the change in amplitude was an independent predictor of the change in BNP levels (R = 0.717, P < 0.01). A 5-month cardiac rehabilitation program improves exercise oscillatory ventilation in chronic heart failure patients by reducing the oscillation amplitude. This effect is associated with a reduction of plasma BNP levels, potentially contributing to an improvement of heart failure.