An algorithm to predict the composition of n -paraffin mixtures from calorimetry is proposed and applied for two thermodynamic approaches: one considering the formation of solid solutions and the other, known as multisolid phase model, assuming the presence of several independent pure solid phases. This methodology compares modeled calorimetric curves with DSC results through Particle Swarm Optimization. In order to better predict the mixtures compositions an appropriate proposal of the objective function was necessary. The present paper elucidates the process of defining an enhanced objective function aiming to improve the quality of predictions allowing the employment of calorimetric curves without information about sample's mass. The deviations in calculated compositions are related to a difference of up to 1 K in the location of calorimetric peaks for solid solution approach. (C) 2017 Elsevier B.V. All rights reserved.