The aim of an optimal design of electrical machines requires the accurate prediction of iron losses for various operating points. For this purpose different iron-loss models have been proposed which intent to describe the loss inducing effects. The most used iron-loss prediction formulas are either physically based, but nevertheless only valid for linear material behavior at low frequencies and low magnetic flux densities, or grounded on a pure mathematical description of the material behavior, that is not more than interpolated measurements. This paper presents a modified loss equation with semi-physically based parameters as well as a first try to explain the nonlinear loss component.