Response surface methodology (RSM) based on a five-level, three-variable central composite design (CCD) was employed for modeling and optimizing the conversion yield of the enzymatic acylation of hesperidin with decanoic acid using immobilized Candida antarctica lipase B (CALB) in a two-phase system containing [bmim]BF(4). The three variables studied (molar ratio of hesperidin to decanoic acid, [bmim]BF(4)/acetone ratio and lipase concentration) significantly affected the conversion yield of acylated hesperidin derivative. Verification experiments confirmed the validity of the predicted model. The lipase showed higher conversion degree in a two-phase system using [bmim]BF(4) and acetone compared to that in pure acetone. Under the optimal reaction conditions carried out in a single-step biocatalytic process when the water content was kept lower than 200 ppm, the maximum acylation yield was 53.6%.