Centella asiatica (CA), known for its health-promoting properties, is rich in bioactive compounds. This study optimised ultrasound-assisted extraction (UAE) parameters to maximise total phenolic content (TPC) and total flavonoid content (TFC) using the response surface methodology (RSM). Ethanol concentration and solvent volume significantly influenced TPC and TFC yields (p < 0.0001), while ultrasonic power had nonsignificant effects (p < 0.05). Time showed no significant effect on TPC (p > 0.05) but influenced TFC due to flavonoids’ sensitivity to degradation (p < 0.05). Variable interactions were negligible (p > 0.05). The relationship between responses (TPC and TFC) and independent parameters could be expressed as the quadratic models fitted with a Predicted R² of 0.8263 for TPC and 0.9006 for TFC. Based on RSM, the optimal conditions—75% ethanol concentration, 87.5 W ultrasonic power, 30 min extraction time, and 20 mL solvent volume—yielded TPC and TFC values of 52.29 ± 1.65 mg/g and 43.71 ± 1.92 mg/g, closely aligning with model predictions at 95% confidence. Additionally, the optimal UAE condition provided asiaticoside of 37.56 ± 4.25 mg/g and madecassoside of 16.91 ± 1.28 mg/g. This study offers valuable insights into the factors influencing UAE efficiency, sustainability, and scalability for recovering bioactive compounds, underscoring its potential as a sustainable method for developing functional food ingredients from CA.