The routine method of monitoring Zircaloy cladding corrosion in nuclear fuel pools is based on eddy current (EC) measurements at a single high frequency in the range of 1-3 MHz. At this frequency the rf wave does not penetrate through the cladding wall and, ideally, the EC response can be correlated to the thickness of the oxide layer that separates the sensor head from the metal substrate. In practice, however, the cladding corrosion is often overestimated by this method due to the primary circuit corrosion products (or crud) that are deposited on the fuel rod surface. In addition, the crud, which is primarily nickel ferrite oxides, may significantly interfere with the EC response due to its ferromagnetic nature. We describe a two-frequency method with four-dimensional vector analyses of the EC response to more precisely assess the cladding corrosion. Two independent approaches for measuring the corrosion damage are suggested in this study. First, the four-dimensional data treatment enables a direct measure of the oxide thickness, even in the presence of ferromagnetic crud. Second, it can also provide a direct measure of the substrate wall thickness, and hence the degree of corrosion, provided the original wall thickness is known. By comparison, the approaches to crud correction currently available prove inadequate, especially if both the thickness and permeability of the crud deposits vary over the fuel rod surface, as is generally the case. The new method has been applied to inactive samples with and without the ferromagnetic crud effect. In the reported laboratory simulations, the oxide thickness and wall thinning were measured independently to within ±4 m, irrespective of unknown crud thickness and permeability.