The loss of surfactant can reduce the technical and economic efficiency of chemical surfactant flooding in the recovery of residual oil. Retention of surfactant molecules is thus recognized as a fundamental problem in chemical recovery based methods. Therefore, reliable methods which are able to rapidly estimate the surfactant retention are of importance. In this communication, a new model is proposed for the determination of surfactant retention in porous media during chemical flooding. The mathematical algorithm adopted in the development of the model is gene expression programming (GEP). The input parameters for the new model are temperature, maximum effluent pH, reservoir rock type i.e., carbonated or sandstone, co-solvent concentration, average molecular weight of surfactant mixture, total acid number (TAN), absolute permeability, mobility ratio, and salinity of polymer. Several statistical and graphical error analyses were applied to assess the performance and accuracy of the proposed model. A comparison was also performed between the newly developed model, a smart method, and a previously published empirical correlation available in literature. The newly developed model performs, overall, superior to the methods compared. Estimations were found to be within acceptable agreement with the literature-reported data of surfactant retention, with an average absolute relative deviation of approximately 16.6%, and a R-squared value of 0.92. (C) 2015 Elsevier Ltd. All rights reserved.