We investigated the performance of a T-type microchannel for mercaptan extraction from light straightrun naphtha (LSRN) with sodium hydroxide solution. The aim of this work is to introduce the microfluidic system as a potential tool for mercaptan extraction from light petroleum products. Modeling the extraction process of mercaptan from LSRN has not been carried out previously. In this regard, mercaptan extraction was modeled by response surface methodology (RSM) and artificial neural network (ANN) to analyze the effect of operating parameters on the mercaptan extraction process. The independent variables are considered as temperature, sodium hydroxide concentration, and the volume ratio of sodium hydroxide to LSRN. Two models were compared based on error analysis of the predicted data. Root mean square error, mean relative error, and determination coefficient for the neural network were 0.5650, 0.4341, and 0.9862, respectively. The values of these parameters for the RSM model were 0.6854, 0.7648, and 0.9798. The results showed that the prediction accuracy for both models is appropriate, but the precision of the neural network model is slightly higher than that of the RSM model. The genetic algorithm (GA) technique determined the optimal values of the independent variables with the aim of maximizing the extraction percentage. The mercaptan extraction percentage value of 85.08% was achieved at 303.15 K, the sodium hydroxide concentration of 20 wt%, and the volume ratio of sodium hydroxide to LSRN of 0.128. Furthermore, results showed a higher mercaptan extraction percentage of the microfluidic system compared to a conventional extractor at the same process condition.