The back-propagation artificial neural network and response surface methodology were used to investigate the estimation capabilities of these two methodologies and optimize the acceptability of desirability functions methodology in an air drying process. The independent factors were the air temperature, air velocity and drying time in the drying process for lavender leaves, while the moisture content, drying rate, energy efficiency and exergy efficiency were selected as the dependent variables or responses. In addition to this isoresponse contour plots were help-ful to predict the results. The artificial neural network models determined an optimum point set at the air temperature equal 46.8°C, the air velocity equal 0.726 m s-1 and the drying time equal 9.72 h to minimize the moisture content and to maximize the drying rate, the energy and exergy efficiencies. At the optimum point the moisture content, drying rate, energy and exergy efficiencies were found to be 0.32 g g-1, 0.29 g g-1h-1, 0.67 and 0.80, respectively.