The goal of this study was to compare the extraction of rosmarinic acid from Melissa officinalis L. using three techniques (heat- -, microwave- and ultrasound- assisted extraction). In order to obtain the conditions that maximize the rosmarinic acid extraction, a response surface methodology was applied using the circumscribed central composite design of three variables with five levels. The relevant independent variables used for the process optimization were time, temperature and ethanol-water proportion for heat-assisted- and microwave-extration, whereas for the ultrasound method the ultrasonic power was variable. The responses used as criteria were the amount of rosmarinic acid was determined by HPLC-DADand the extraction yield of the obtained residue. Ultrasound extraction proved to be the most effective method, capable of yielding 86.3 +/- 4.1 mg rosmarinic acid/g plant per dry weight (dw) at the optimal extraction conditions (33.0 +/- 3.2 min, 371.7 +/- 193 W and 39.9 +/- 1.4% of ethanol). According to the content of rosmarinic acid, microwave- and heat-assisted extractions techniques were less effective, producing 49.4 +/- 2.3 (at 26.5 +/- 2.1 min, 108.6 +/- 10.2 degrees C and 25.5 +/- 0.9% of ethanol) and 59.4 +/- 2.2 (at 106.2 +/- 5.1 min, 88.0 +/- 2.9 degrees C and 34.5 +/- 1.6% of ethanol), respectively. Additionally, the solid/liquid ratio effect at the optimal values in a dose-response format was studied in view of its plausible transference at industrial level, showing a decreasing non-linear pattern from 5 to 120 g/L. In brief, the obtained results highlight the potential applications of using the leaves from M. officinalis as a source of rosmarinic acid. (C) 2017 Elsevier B.V. All rights reserved.