The objective of the present study is to identify the most suitable corrosion degradation model, fitted with real corrosion depth measurement data sets and to reproduce the corroded steel plate surface as a function of time and spatial distribution using advanced statistical methods. An approach for adequately identifying the best-fitted model to real corrosion depth measurement data sets is employed. Two distinct statistical methods for generating a statistical representation of a corroded plate surface in the case of significant and insignificant correlation of the corrosion degradation are provided. A sequence-dependent data analysis is performed based on the fast Fourier transform, which is used as an input for a random field modeling of corroded steel plate surfaces. The output of this study represents very important information about the corroded plate surface topology that can be used in any advanced finite element analyses of structural integrity assessment. The formulations can be adapted to any structural components and corrosion environments.