<p>This article presents an effective load estimation strategy that pairs an analytical model with a dynamic Kalman estimator for Inductive Wireless Power Transfer (IWPT) system that eliminates feedback from the secondary-side to the primary-side. Given that sensed primary-side current and voltage quantities are vulnerable to noise, conventional approaches tend to propagate this noise, thereby impacting the accuracy and tracking of the identified load parameters. Therefore, this article proposes the implementation of a two-step strategy for identification under varying output load conditions. First, the output load resistance is estimated analytically as a baseline parameter. Then, based on the recognized resistance parameter, a state-space dynamic Kalman model of the system is derived to track the output load voltage. This two-step identification strategy has the advantage of robust estimation in the presence of noise and load variations as it decouples the nonlinear system via a linear estimator. The efficacy of the proposed identification approach is confirmed through simulation and experimental validation using a 75 W IWPT hardware test-bench. Under dynamic operation, the proposed two-step identification approach reduces the estimation error by 32% at maximum variation range.</p>