Scale-up of high performance liquid chromatography (HPLC) using mathematical models has been found in the literature for separation of binary mixtures or monocomponent materials. This procedure provides graphical representation, in contrast to direct scale-up, which has been usually employed for separation of natural fraction components. In this report, the application of models to the scale-up of isocratic separations, from literature data, of carotenoids, vitamins, ginsenosides, and monoterpenes fractions, in terms of use of larger columns and sample overloads, is discussed. Statistical moment analysis and an ideal rate model were applied using a computer spreadsheet to estimate parameters from analytical data and predict semi-preparative separations. Non-competitive effects between the components were assumed, due to the few available data. The predicted chromatograms showed good agreement with the experimentals, demonstrating the applicability of scale-up using models on separation of natural fractions.