Surface/interfacial tension is an important parameter in gas-water, gas-oil and oil-water systems, which can affect various aspects of underground processes as well as related operations in surface equipment. In this regard, the controlling role of this parameter on the distribution of fluids, their mobility in the porous medium, the trapping of phases and the saturation of the remaining fluids in the underground layers is undeniable. Therefore, it is considered as one of the inseparable factors in the design and simulation of gas storage and reproduction methods in natural production from oil and gas reservoirs and EOR. In addition, surface/interfacial tension plays an important role in phase separation and emulsion formation in fluid transmission lines and equipment of the production unit. Thus, the modeling of this important parameter can lead to a deeper understanding as well as better simulation of the mentioned processes. Despite the laboratory measurement of this parameter, it is necessary to predict or estimate it in different conditions as part of simulating the related processes. In this research, the modeling porcesses of surface/interfacial tension has been addressed using new intelligent (artificial intelligence) and hybrid methods for three systems composinig gas-water, gas-oil and oil-water interfaces. The aforementioned intelligent methods include LS-Boost, MLP, LS-SVM, CMIS, RBF and ELM. Also, some of these smart methods are combined with optimization algorithms such as ICA, GWO, WOA, LM, BR, SCG and CSA in order to adjust parameters or hyper-parameters. In this way, the consistency of data predicted by models against the laboratory measurements is checked using various graphical and statistical analysis. The results of the modeling strategies shows that the LS-Boost and CMIS based algorithms had the best performance among the developed models in this research. In fact, the values of R2 for CMIS, GWO-LSBoost and WOA-LSBoost are 0.9963, 0.9991 and 0.
