<p>This work revisits the Kalman filtering topic to propose a new idea of implementation for the data assimilation step (also called the update step). Among the conventional nonlinear algorithms, the unscented Kalman filter (UKF) is a reference since it employs unscented transformation rather than analytical linearization. Moreover, its update step is executed at once by fixing the state forecasts over the approximations. Therefore, we take the UKF as a basis to derive the iterative UKF (UKF-I), which incorporates measurement information individually in order to refine the state forecasts and thereby provide better approximations for the posterior data assimilation. This specific modification is capable of potentiating the filter, yielding improved estimation accuracy and convergence. We also derive the worst-case complexity order of each algorithm to raise additional discussions. Two practical examples are provided to illustrate the effectiveness of our proposal under challenging scenarios. The numerical results corroborate the benefits of better accuracy and convergence speed with the UKF-I.</p>