Intraoperative hypotension, which frequently occurs during surgery, can lead to unfavorable results. However, general blood pressure monitoring has limitations in continuously detecting hypotension in a patient, and there is a risk of not recognizing the presence of hypotension. Even short-term hypotension can damage major organs. The hypotension prediction index (HPI) indicates the likelihood that a patient will develop hypotension within the next 5 to 15 minutes in a hemodynamically stable state without hypotension. The HPI algorithm was developed to calculate the probability of hypotension using a machine-learning algorithm to characterize continuous arterial pressure waveforms in 1,334 surgical and critically ill patients. HPI provides parameters for determining the cause of hypotension (preload, contractility, or afterload), such as stroke volume variability, dP/dt max , and dynamic arterial elastance. Although there is still no standard protocol for preventing or treating hypotension using HPI values, predicting the risk continuously and the cause of hypotension during surgery in real-time may help prevent harmful hypotension during surgery. In this review, we introduce the concept of HPI and discuss whether it can be applied in actual clinical practice.