Umbilical artery (UA) blood velocity pulsatility is a widely used biomarker for placental pathology and is associated with intra-uterine growth restriction and still-birth in human pregnancy. It has been hypothesized that reflected waves originating from the placenta cause the association between placental pathology and modified UA blood velocity pulsatility. We present a new method to measure these wave reflections, anticipating they can yield more specific information about pathology than pulsatility.A forward and reflected wave was determined from the analysis of blood velocity waveforms obtained by Doppler ultrasound at three locations along the length of the UA in 148 pregnant volunteers. 128 of the 192 measurement triplets with suitable data quality were consistent with the wave reflection model. This analysis method was validated using Monte-Carlo simulations based on a set of three velocity waveforms that were consistent with the wave reflection model. In the absence of measurement noise, error in the calculated reflected velocity waveform is negligible. Measurement noise degrades the precision of these estimates such that for noise typical of clinical ultrasound, the first-harmonic reflection coefficient could be estimated to within 2%.