Histograms are commonly used to measure heart rate variability (HRV), particularly for measurements over a 24-hour period. In this study, a five-class density histogram (FCDH) is proposed to extend the application of histograms for short-term HRV evaluation. The cumulative density, an index derived from the FCDH, is compared with conventional time- and frequency-domain measures, namely (1) low-frequency power (LF); (2) high-frequency power (HF); (3) the power ratio (LF/HF); (4) triangular interpolation of NN intervals (TINN); and (5) the HRV triangular index (HRVi). A total of two hundred subjects participated in the HRV assessment. They came from four equal-sized groups: the healthy young (Young), the healthy old (Old), patients with coronary artery disease, and patients with congestive heart failure. After 7 days and 14 days, 46 subjects from the Young and Old groups were retested following the same protocol in order to examine the reliability of the FCDH. Results show significant differences between the four groups for the indices of cumulative density, LF, HF, TINN, and HRVi (p < 0.001), but not for LF/HF (p = 0.269). The cumulative density shows good-to-excellent interclass correlation (ICC) with an average ICC = 0.783 (95% confidence interval 0.676-0.864), which is higher than the average ICCs of LF (0.695), LF/HF (0.669), TINN (0.686), and HRVi (0.615), but lower than that of HF (0.893). The cumulative density derived from the FCDH is thus a valid and reliable method for short-term HRV assessment.