Objective To explore the effect of self-assessment of pain in perioperative pain management of total knee arthroplasty (TKA). Methods A total of 140 patients undergoing TKA from March 2016 to March 2017 were randomly divided into the control group and the trial group. The patients in the two groups were received the same education relating to pain knowledge. The intensity of pain was assessed by nurses in the control group, while in the trial group, it was assessed by patients themselves. According to the assessment of pain, treatments were given to both groups. Time of pain assessment, types and frequencies of temporary rescue medicine, pain intensity, the score of Self-efficacy for Rehabilitation Outcome Scale (SER) and the range of motion (ROM) of knee were observed and recorded. Results There were 132 patients who completed the final observation, with 67 in the trial group and 65 in the control group. There were significant differences between the two groups in evaluation time of pain (t=–2.736, P=0.007), types and frequencies of temporary rescue medicine (χ2=10.276, P<0.05), the overall postoperative pain score (Z=–2.146, P=0.032), average hospitalization time after surgery (t=–2.468, P=0.015), SER scores 7 days after surgery (F=2.390, P=0.018) and 14 days after surgery (F=3.427, P=0.001), and ROM at the postoperative day 7 (F=2.109, P=0.037); there were no significant differences in postoperative daily pain scores (Z=–1.779, P=0.077), SER scores at the postoperative day 3 (F=1.010, P=0.314), ROM at the postoperative day 1 (F=1.319, P=0.189) and day 14 (F=1.603, P=0.111). Conclusion Self-assessment of pain can motivate TKA patients to take part in pain management, and more accurate response to the pain intensity will help to optimize the management of perioperative pain and reduce the workload of the health staff, thereby contributing to enhanced recovery.