In the last decades, response surface methodology (RSM) and artificial neural network (ANN) has become the most preferred methods for non-parametric modeling and optimization of separation processes in chemical engineering. This paper presents a comparative study between RSM and ANN for reactive extraction optimization. Reactive extraction of tartaric acid from aqueous solution using Amberlite LA-2 (amine) has been chosen as the case study. The extraction efficiency was modeled and optimized as a function of three input variables, i.e. tartaric acid concentration in aqueous phase C_AT (g/L), pH of aqueous solution and amine concentration in organic phase C_A/O (% v/v). Both methodologies have been compared for their modeling and optimization abilities. According to analysis of variance (ANOVA) the coefficient of multiple determination of 0.841 was obtained for RSM and 0.974 for ANN. The optimal conditions offered by RSM and genetic algorithm (GA) have led to an experimental extraction efficiency of 83.06%. On the other hand, the ANN model coupled with GA has conducted to an experimental reactive extraction efficiency of 96.08% for the following optimal conditions of ^*C_AT = 5.58 g/L; ^*pH 1.84 and ^*C_A/O = 6.99 (% v/v). The value of 96.08% is the maximal value of extraction efficiency obtained throughout this work. Thus, ANN-GA has demonstrated the ability to overcome the limitation of quadratic polynomial model in solving optimization problem for this case study. Both models have been employed for construction of response/output surface plots in order to reveal the influence of input variables on extraction efficiency as well as to figure out the interaction effects between variables.