Abstarct: Abstract In this study, the solubility of carbon dioxide in different ionic liquids has been predicted using the group method of data handing neural network (GMDH) for a wide range of temperature and pressure. The selected descxriptors include temperature and pressure as well as molecular weight, melting point and density of ionic liquids. The data set consists of 1795 experimental data points of CO2 solubility in 19 ionic liquids for a wide range of temperatures and pressures. The solubility data of carbon dioxide in 15 ionic liquids were divided into two groups: training and internal test sets, and the solubility of carbon dioxide in 4 other ionic liquids were considered as the external test set. After training and optimization of the GMDH network parameters, the performance of the optimized model was investigated by the internal and external test sets. The results showed that the optimized GMDH model can predict the solubility of carbon dioxide in ionic liquids with acceptable accuracy for a wide range of temperatures and pressures. The root mean square error (RMSE) and determination coefficient (R2) were obtained for the internal test set 0.0609 and 0.9673 and those for external test set 0.0663 and 0.9677, respectively.