Deceleration capacity (DC) of heart rate, a novel indicator of autonomic nervous system, is calculated from the unevenly sampled RR tachogram. We aim to study whether the nature of uneven sampling of the RR tachogram affects the performance of DC. Interpolation and resampling procedures were applied to the RR tachogram to generate an evenly sampled RR interval (RRI) time series. In particular, the resampling frequency, the interpolating approach, and other relevant parameters were examined for optimizing the performance of the modified DC. DC calculated from the resampled RRI time series was compared to the conventional one computed from the raw RR tachogram. The conventional and modified DCs were applied in distinguishing the chronic heart failure (CHF) patients and healthy subjects. The results show that the modification improved the diagnostic accuracy of CHF from 85.11% to 90.43%, suggesting that the resampling procedure enhances DC in the assessment of CHF.