Surfactant flooding is known as a tertiary oil recovery technique for increasing recovery factor from depleted sandstone and carbonate oil reservoirs. The currently used industrial surfactants are too expensive and they have detrimental environmental effects; therefore, natural surfactants have been proposed by many researchers to overcome these challenges. Additionally, adsorption of surfactants onto solid surfaces makes this oil recovery method less effective for enhanced oil recovery applications. Therefore, it is crucial to develop a new rapid, simple, and accurate model. In this study, least square support vector machine (LSSVM) optimized with coupled simulated annealing algorithm (CSA) is used for accurate prediction of kinetic adsorption of natural surfactants on both carbonate and sandstone rocks. Predicted values by this model were in an excellent agreement with actual values with a coefficient of determination of 0.990. The results demonstrated that the proposed LSSVM-CSA model has the best performance in comparison with the other well-known kinetic models. Furthermore, the effect of surfactants adsorption on the properties of the produced oil was investigated. It was shown that in systems where surfactants have less adsorption, the produced oil has higher density and viscosity.