Abstract: Objective To study the effect of multimodal analgesia in pain management after unilateral total knee arthroplasty. Methods Ninety osteoarthritis patients who underwent unilateral TKA in our hospital from September 2014 to August 2016 were included in this study.They were divided into three groups with different analgesia regimens with 30 patients in each group.Group A: preoperative oral celecoxib and tramadol hydrochloride + periarticular "cocktail" injection during operation + patient controlled analgesia pump after surgery.Group B: periarticular “cocktail” injection during operation + patient controlled analgesia pump after surgery without preoperative drugs.Group C: patient controlled analgesia pump after surgery without preoperative drugs and periarticular "cocktail" injection during operation.Visual analogue scale scores and range of motion were selected as indicators of effects at several different time points after the surgery. Results No signi ficant difference was found in the demographic data of the three groups (P＞ 0.05).The VAS scores were recorded as A(2.5±0.6)＜ B(2.8±0.6)＜ C(3.0±0.7)at 6 h, A(3.4±1.1)＜ B(3.8±0.8)＜ C(4.2±1.4) at 12 h, A(4.5±1.2)＜ B(4.9±1.3)＜ C(5.4±0.9) at 24 h, and A(5.6±1.4)＜ B(6.1±1.3)＜ C(6.3±1.4) at 48 h after TKA (P＜ 0.05, respectively).No signi ficant difference was found between the three groups at 72 h and 7 d after TKA (P＞ 0.05).The ROM were recorded as A(46.4±5.4)＞ B(42.5±7.3)＞ C(40.4±5.2) at 24 h, A(58.6±4.2)＞ B(56.2±4.7)＞ C(53.3±6.1) at 48 h, A(64.6±5.2)＞ B(62.5±8.8)＞ C(60.3±7.4) at 72 h after TKA (P＜ 0.05, respectively).No signi ficant difference was found between the three groups at 7 d. Conclusion The study shows that multimodal analgesia bene fits to pain management and achieve better ROM in early stage after TKA.