The ethanolysis of sunflower oil catalyzed by calcium oxide was modeled and optimized in terms of the following operating conditions: reaction temperature (65 similar to 75 degrees C), ethanol:oil molar ratio (6:1-18:1), catalyst loading (10-20% based on oil weight) and reaction time (360-480 min). Response surface methodology (RSM) and artificial neural network (ANN) approaches were used for modeling the content of fatty acid ethyl esters (FAEE) and optimizing the four process variables. Both models were determined to be reliable in terms of predicting the FAEE content, but the ANN model was found to be more accurate than the RSM model. The highest FAEE content of 99.2% was determined using the ANN model combined with a genetic algorithm optimization method, which agreed well with the experimental value (97.8%). A good agreement between the predicted and actual maximum FAEE contents was observed for both models. The generalization of the ANN model developed for heterogeneously catalyzed alcoholysis was also tested on several oily feedstocks. (C) 2015 Elsevier Ltd. All rights reserved.