Evaluating the high-frequency winding loss accurately is crucial for the design of modern high-frequency power converters. This paper proposes a novel experimental method to accurately measure the in-situ inductor winding loss, which separates out the winding loss from the core loss through the reactive voltage cancellation concept. The proposed in-situ measurement can account for the complete winding loss including impacts from non-ideal field distributions by testing the inductor with the core attached, e.g., the winding edge effect, bypass flux, fringing flux, and the non-linear dynamic behaviour of the core, which cannot be well modelled with the existing analytical or simulation methods. This method has immunity against the probe phase discrepancy error, since it is designed to measure a pair of voltage and current that are in phase. This approach can be considered as the first attempt of applying the reactive voltage cancellation concept in measuring winding loss, while this concept was originally brought up for core loss and applied for core loss measurement only. By performing a Triple Pulse Test (TPT) procedure, the winding losses under practical large signals with dc-bias and rectangular voltage can be evaluated with the proposed testing circuit. The proposed approach is compared and verified against the conventional methods relying on (1) small-signal impedance measurements and FFT analysis (2) in-situ measurement with the two winding method to exclude the core loss. The presented method provides a foundation for the accurate in-situ evaluation of winding loss covering all the large-signal and non-linear effects.