Optimization of the supercritical fluid extraction of Quercus cerris cork was carried out using Box-Behnken design of experiments and response surface methodology (RSM). The optimized process variables were temperature (T: 40, 50 and 60 °C), ethanol content (EtOH: 0.0, 2.5 and 5.0 wt%) and CO2 flow rate (QCO2: 5, 8 and 11 gmin-1). The studied responses were total extraction yield (ηTotal), friedelin concentration of the extract (CFriedelin), and selectivity towards friedelin (αF,nF). The linear effect of EtOH was by far the most influent operating condition (Pareto analysis) and the highest yield (ηTotal = 2.2 wt%) was attained with 5.0 wt% EtOH. The RSM model estimates maximum friedelin concentration in the extracts (38.2 wt%) to occur without cosolvent (0 wt% EtOH) for the lowest T (40 °C) and QCO2 (5 gCO2min-1). As for selectivity, the experimental αF,nF values were always higher than 1.0 and reached 3.1 (at 50 °C, 5 wt% EtOH, 11 gCO2min-1). Altogether, these results suggest friedelin can be selectively removed from Quercus cerris cork by supercritical fluid extraction within the range of experimental conditions studied.