Abstract Response surface methodology (RSM) was used to optimize ultrasound-assisted extraction (UAE) of functional components from perilla leaves. The factors investigated were ethanol concentration, extraction temperature, and extraction time. The results revealed that ethanol concentration had significant effects on all extraction parameters. Based on the RSM results, the optimal conditions were an ethanol concentration of 56%, a UAE temperature of 54 °C, and a UAE time of 55 min. Under these conditions, the experimental TPC (total phenolic content), RA (rosmarinic acid), FRAP (ferric reducing antioxidant power) and DPPH (1,1-diphenyl-2-picrylhydrazyl) values were 48.85 mg GAE/g DW (mg gallic acid equivalent /g of dry weight), 31.02 mg/g DW, 85.55 μmol Fe2+/g DW and 73.35%, respectively. The experimental values were in agreement with those predicted by RSM models, confirming suitability of the model employed and the success of RSM for optimization of the extraction conditions.