It has been reported that increased intermittent non-Gaussian fluctuations in the instantaneous amplitude of low-frequency heart rate variability (LF-HRV) are related to high mortality risk in cardiac patients. However, little is known about the physiological origin of the amplitude modulation of LF-HRV. The purpose of this study was to clarify the relationship between amplitude modulation of LF-HRV and that of low-frequency blood pressure variability (LF-BPV). Eight normal male subjects performed movie-watching and calculation tasks in a sitting position for 40 min each while electrocardiogram and continuous blood pressure waveforms were recorded. From these signals, we calculated the instantaneous amplitude of the LF-band RR interval (RRILFamp) signal and that of the LF-band systolic blood pressure (SBP) signal (SBPLFamp) via band-pass filter and Hilbert transform. All subjects exhibited significant and relatively high positive correlation coefficients between RRILFamp and SBPLFamp in both tasks (mean Pearson correlation coefficient > 0.45). Mean coherence in the 0.01–0.05 Hz band between RRILFamp and SBPLFamp was also significant in all but one subject (mean coherence > 0.42). These results indicate a relatively high positive correlation between the amplitude modulation of LF-HRV and that of LF-BPV. We calculated the peak time lags of the cross correlation between RRILFamp and SBPLFamp in the 0.01–0.05 Hz band. A negative peak time lag implies that the amplitude modulation of LF-HRV precedes that of LF-BPV. All subjects exhibited negative peak time lag in the movie-watching task. All but one subject exhibited negative or zero peak time lag in the calculation task. These results imply that the amplitude modulation of LF-HRV precedes that of LF-BPV in the frequency range of 0.01–0.05 Hz.