The methanolysis of hempseed oil catalyzed by potassium hydroxide was studied using the full central composite rotatable design. The methanolysis reaction was modeled and optimized by the response surface methodology (RSM) and an artificial neural network model combined with a genetic algorithm (ANN-GA). All individual factors have significant influence on the FAME content (p-value < 0.05). Based on the coefficient of determination and the mean relative percentage deviation (0.996 and 0.2% for RSM and 0.870 and 2.9% for ANN, respectively), the RSM model demonstrated a better data modeling and more efficient prediction capability than the ANN-GA model. Also, the former predicted the more favorable optimum reaction temperature and methanol:hempseed oil molar ratio (43.4 degrees C and 6.4:1) than the latter (56.8 degrees C and 8.5:1) at a somewhat higher catalyst loading (1.2% versus 1.0%). Moreover, the RSM model predicted the maximum FAME content of 99.8% (actual content: 98.5%) in 30 min, while the ANN-GA model resulted in the best predicted yield of 97.1% (actual content: 97.5%) within 12 min. Generally, the properties of obtained FAME are within the specifications of the biodiesel standards. (C) 2015 Elsevier Ltd. All rights reserved.