Amongst the present water treatment technologies membrane separation is gaining focus nowadays. Emulsion liquid membrane is one of the advanced strategies used for pollution removal from waste water in which the use of biosurfactant makes this approach eco-friendly technique. The norfloxacin belongs to the fluoroquinolones class of antibiotic and extensively used for urinary tract infections, which is the model pollutant in this study. Saponin was solvent extracted from soapnuts and used as biosurfactant. The main and interactive effects of parameters such as saponin concentration (0.01-0.03 g/100 mL), NaOH concentration (0.25-0.75 M) and Initial Norfloxacin concentration (25-75 mg/L) were investigated using Response Surface Methodology (RSM)-Box Behnken Design (BBD) and Artificial Neural Networks (ANN) design. The results suggested that developed 3-8-1 ANN model evaluated in terms of performance measure (R 2 : 0.9806, Chi square: 0.02) and error functions (RMSE: 0.11, MAE: 0.09, SEP: 1.16, MPE: 1.76) demonstrated superiority more precisely than the RSM model in terms of both data fitting and prediction capability. The optimal conditions for ELM were found to be: initial norfloxacin concentration 67.65 mg/L, saponin concentration 0.021 mg/100 mL, and NaOH concentration 0.36 M and the extraction efficiency at these parameter settings would be 91.27%.