The relation between changes in blood pressure and changes in autonomic activity over a very short period of time has not been reported thus far. To examine this relation, we here introduced a new method of power spectrum analysis with wavelet transformation, which has very fine time resolution and is able to assess changes in autonomic activity quantitatively even during movement. Our subjects were 15 hypertensive and 17 normotensive subjects. A head-up tilt test was performed in all subjects, and during the test, electrocardiogram and blood pressure were recorded continuously. The power spectrums for both parameters were calculated simultaneously every 5 s using wavelet transformation. The high frequency of the RR interval of the electrocardiogram (RR-HF) and low frequency of systolic blood pressure (SBP-LF) were defined and calculated as markers of parasympathetic and alpha-1 receptor blocker, bunazosin-sensitive sympathetic activity, respectively. Focusing on the changes for 2 min immediately after head-up tilting, it was found that the changes in SBP-LF and RR-HF were significantly delayed, by at least 40 s, in hypertensives compared with normotensives and also in elderly compared with non-elderly subjects. Multiple regression analysis demonstrated that the instantaneous change in RR-HF was the most important confounding factor for a fall in blood pressure immediately after head-up tilting. In conclusion, real-time changes in autonomic activity calculated by wavelet transformation may provide sensitive and useful information about acute changes in cardiovascular regulation, such as delayed reaction of the autonomic regulation after head-up tilting, that may be major causes of the blood pressure fall in hypertensive and elderly subjects.