In this work, neural networks were used to prediction of phenol extraction using TBP solvent. Volume ratio of organic to aqueous, rotor speed, temperature, pH and time were selected input variables of the networks whereas extraction efficiency was selected output variable. The network was trained and evaluated using 184 experimental data. The network with one hidden layer with 10 neurons was selected as optimal network. RMSE and R^2 are the criteria for evaluating the performance of the network were obtained to 1.0393 and 0.99, respectively. Optimal process conditions, the ratio of organic to aqueous 0.22, rotor speed of 350 rpm, temperature 22.76 °C, pH 7.5 and time of agitation 15.86 minutes and the extraction efficiency corresponding to these conditions 96.23 were obtained.
)