Using two entropy-based measures, namely the approximate entropy and sample entropy measures, we studied the complexity of heart rate variability signals obtained from professional shooting athletes in the situations of rest and practice match. The results demonstrate that the values of two measures calculated from the resting signals are both greater than those calculated from the training signals, which means that the signals collected during the match are more regular compared to those acquired in a resting state. For a better application of the two methods, we further investigated the influences of two factors: threshold r and data length N, on the performance of the algorithms. Although both approaches have the ability to discriminate the complexity of heart beat interval series from different states of the shooters, provided that the parameters required by the algorithms are chosen within a proper range, it still seems that sample entropy method is more appropriate in quantifying the short-term heart rate variability signals for shooting athletes, especially when the time series are only several hundred points long.