Cistus ladanifer L., Acacia dealbata L., and Aloysia citrodora Paláu were subject to an optimization procedure for two extraction techniques (heat-assisted extraction (HAE) and ultrasound-assisted extraction (UAE)). The extracts were then analyzed by HPLC-DAD-ESI/MS for their phenolic profile (cistus—15 compounds, acacia—21 compounds, and lemon verbena—9 compounds). The response surface methodology was applied, considering four varying factors: ethanol percentage; extraction time; temperature/power; and S/L ratio, generating two responses (the major phenolic compound, or family of compounds, and the extraction yield). For cistus, both techniques optimized the extraction yield of punicalagins, with UAE proving to be the most efficient extraction method (3.22% ethanol, 22 min, 171 W, and 35 g/L). For acacia, HAE maximized the extraction of procyanidin (74% ethanol, 86 min, 24 °C, and 50 g/L), and UAE maximized the content of myricetin (65% ethanol, 8 min, 50 W, and 50 g/L). For lemon verbena, HAE favored the extraction of martynoside (13% ethanol, 96 min, 49 °C and 17 g/L) and forsythiaside UAE (94% ethanol, 25 min, 399 W, and 29 g/L). The optimal conditions for the extraction of compounds with high added value and potential for use in pharmaceuticals and nutraceuticals were defined.