Physiologic systems generate complex fluctuations in their output signals that reflect the underlying dynamics. In order to detect the effect of circadian rhythm for heart rate variability signals, we apply base-scale entropy method and power spectral analysis to the 24-hour heart rate variability signals. The results show that 1) such profound circadian- and pathologic-dependent changes are accompanied by changes in base-scale entropy and power spectral distribution, but by little changes in approximate entropy; 2) the circadian regulating ability of vagal nerve is clearly decreasing for congestive heart failure subjects; 3) the base-scale entropy is more sensitive than spectral analysis method to distinguishing wake/sleep states and identifying patterns generated from healthy and pathologic states, meanwhile, the base-scale entropy changes reflect corresponding changes in autonomic nerve outflow. With the suppression of vagal tone and dominance of sympathetic tone in congestive heart failure subjects, there are more variabilities in the m-words form π due to the trends in the data. So the higher base-scale entropy belongs to congestive heart failure subjects. With the decrease of sympathetic tone during sleep, the base-scale entropy drops in both healthy and congestive heart failure subjects. Finally, in order to further investigate the effect of series length, we calculate the base-scale entropy for different length series and find that the series length nearly has no influence on the result.