According to several studies, the citrus fruits waste material has a high amount of bioactive compounds such as citric acid, a strong antimicrobial agent. The objective of this work was to optimize the ultrasound-assisted extraction of citric acid from orange, lemon and lime peels using the response surface methodology (RSM). For the extraction optimization, a central composite design coupled with RSM was implemented by combining five levels of the variables time (2–45 min), power (50–500 W), and ethanol proportion (0–100%). The levels of citric acid were determined by ultra-fast liquid chromatography coupled to a diode array detector. The highest citric acid contents were obtained from orange peels, followed by lime, with the optimal processing conditions yielding 10 and 8 g per 100 g of dry peels, respectively. In both cases, the extraction was maximized with a low ultrasonic power for a long period of time, using considerably low ethanol proportions. For lemon peels, citric acid was not detected in some runs, so it was not possible to construct a predictive model. This work provided information on the optimal conditions for the extraction of citric acid as a functional agent and contributes to the valorization of citrus biowaste.