OBJECTIVE: To observe and evaluate the effects of transcutaneous electrical acupoint stimulation (TEAS) at different time points on postoperative analgesia in perioperative period in patients undergoing shoulder arthroscopic surgery, and to explore the optimal time to use TEAS for shoulder arthroscopic surgery. METHODS: A total of 120 patients undergoing unilateral shoulder arthroscopy under general anesthesia, graded withⅠtoⅡaccording to ASA criteria were randomly divided into 3 groups, 40 patients in each one. The patients in the group A were treated with preoperative TEAS at sham acupoints combined with postoperative TEAS at Hegu (LI 4) and Neiguan (PC 6); the patients in the group B were treated with preoperative TEAS at Hegu (LI 4) and Neiguan (PC 6) combined with postoperative TEAS at sham acupoints; the patients in the group C were treated with TEAS at sham acupoints before and after operation. The parameters of TEAS were dilatational wave, 2 Hz/100 Hz in frequency, 30 min. When the resting-state visual analogue scale (VAS) of incision was more than 3 points, the patient-controlled intravenous analgesia (PCIA) pump of sufentanil was administered to maintain the VAS no more than 3 points. The time point when PCIA pump was firstly used, the used dose of sufentanil and effective pressing number of PCIA pump within 24 hours after surgery were recorded. Intraoperative anesthetic doses were recorded in the three groups. The resting-state and task-state VAS were evaluated at 0, 6, 12, 24 hours after surgery; the patient's satisfaction rate and adverse effects were recorded. RESULTS: The time when PCIA pump was firstly used in the group A and the group B was significantly longer than that in the group C, and the used dose of sufentanil and effective pressing number of PCIA pump within 24 hours after surgery in the group A and group B were significantly less than those in the group C (all <0.05); the incidence of postoperative nausea-vomiting and sore throat was reduced (all <0.05). The time when PCIA pump was firstly used in the group A was significantly longer than that in the group B, and the used dose of sufentanil and effective pressing number of PCIA pump within 24 hours after surgery in the group A were significantly less than those in the group B (all <0.05); no significant difference of the incidence of postoperative nausea-vomiting and sore throat was observed between the group A and group B (both >0.05). There were no statistically significant difference in VAS score at different postoperative time points and postoperative analgesia satisfaction rate between the two groups (all >0.05). CONCLUSION: Perioperative TEAS could improve the postoperative analgesia in patients undergoing arthroscopic shoulder surgery, delay the time when PCIA pump is firstly used, reduce the dosage of postoperative analgesics and adverse events. Compared before surgery, postoperative TEAS has better analgesia.