The complexity of lipid feedstocks and the lack of data on physical properties hinder the simulation of oleochemical processing units. In this work, an iterative lumping approach is proposed to define an adequate number of key compounds such that diversification between lipid feedstocks becomes possible, while keeping the determination of physical properties as required for process modeling manageable. As a case study, the iterative lumping approach is used for simulation and optimization of a fatty acid distillation plant. For predicting vapor-liquid equilibria of fatty acids, the best results were acquired using the property method universal quasichemical-Hayden-O'Conell. Using the iterative lumping approach, 11 key compounds were selected to represent the feedstock. The process model properly predicts the product composition, yield, purity and heat duty. The most important process parameters are found to be side-reflux-ratio, reboiler-outlet-temperature, and heat-duty of the pitch-distiller. For optimization, an increase of the side-reflux-ratio and reboiler-outlet-temperature is recommended.