It is known that the shape of arterial blood pressure waveform reflects the hemodynamic status, but quantitative analyses of waveform are seldom reported. In this study, the arterial pressure wave was recorded with an intra-arterial catheter and pressure transducer, converted to digital data, and stored in a computer. The data were analyzed by two methods of spectrum analyses, i.e., fast Fourier transform (FFT for both power spectrum and phase information) and the autoregressive model (ARM). Pressure waveforms during normotension were compared with pressure waveforms during hypotension induced by vasodilators (sodium nitroprusside and nitroglycerin) in the human study, and by the vasodilators, hemorrhage or 3% isoflurane in the animal study. Reconstruction of waveforms by inverse FFT was also performed, and compared with the original waveforms. FFT is more useful than ARM for analysis of data with known frequencies, such as respiratory and heart rates. Using inverse FFT, the wave can be reconstructed by the base frequency (heart rate) to 4th harmonic data with an error margin of less than 2%. The amplitude and phase information of heart rate frequency up to its 4th harmonic, and amplitude of respiratory frequency, were found to be the essential parameters to describe the shape of the arterial pressure waveform.