The proper understanding of the cardiovascular mechanisms involved in complaints of short-lasting dizziness and the evaluation of unexplained recurrent syncope requires continuous monitoring of cardiac stroke volume (SV) in addition to blood pressure and heart rate. The primary aim of the present study was to evaluate a pulse wave analysis method that calculates beat-to-beat flow from non-invasive arterial pressure by simulating a non-linear, time-varying model of human aortic input impedance (Modelflow; MF), by comparing MF stroke volume (SV(MF)) to Doppler ultrasound (US) flow velocity SV (SV(US)). A second purpose was to compare the two methods under two different conditions: the supine and head-up tilt (30 degrees ) position. SV(US) and SV(MF) with non-invasive arterial pressure (Finapres) as input to the aortic model were measured beat-to-beat during spontaneous supine breathing and in the passive 30 degrees head-up tilt (HUT30) position in six normotensive healthy humans [three females, mean age 24 (21-26) years]. There were variations in supine SV track between the two methods with zero difference and a SD of the beat-to-beat difference (MF-US) of 4.2%. HUT30 induced a systematic difference of 10.5% and an increase in SD to 6.9%, which was reproducible. Beat-to-beat changes in SV in the supine resting condition were equally well assessed by both methods. Systematic differences appear during HUT30 and show opposite signs. The difference between the two methods upon a change in body position may be attributed to limitations in each method.