Emulsion liquid membrane (ELM) has been shown to be an alternative separation technology for some specific substances from their solutions. In this study, the extraction of L-phenylalanine from sodium chloride solutions by ELM was selected as a model system in which sulfonated kerosene was newly employed as the diluent of membrane phase. Meanwhile, a back propagation neural network (BPNN) improved by genetic algorithm (GA) was first established to simulate the concentration of L-phenylalanine in external phase and extraction efficiency with the time under various operational conditions. The results show that, a stable ELM with an extraction efficiency of 94.4% and a concentration ratio of 106.1 can be acquired. The back propagation neural network improved by genetic algorithm (GABP) is also demonstrated as an effective tool to simulate an ELM process in real-time with average residual errors of 0.0722 g/L and 1.6563% in terms of the concentration of L-phenylalanine in external phase and extraction efficiency, respectively. (C) 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.