Photoplethysmography (PPG) can acquire a variety of biometric information such as pulse rate and SpO2 with a single sensor. If it can automatically detect respiratory information, its value will more increase. Therefore, in this research, we propose a method to extract a respiratory waveform from PPG. Since the systolic wave height of the PPG signal becomes small due to the decrease in the stroke volume during inspiration, our proposed method separates the PPG waveform into heartbeat fluctuation component and respiratory fluctuation component by deriving fluctuation of the systolic wave heights.