Abstract: Objective : To compare the curative effect and adverse reaction of patient -controlled epidural analgesia (PCEA) with that of patient-controlled intravenous analgesia (PCIA). Methods : Sixty patients undergoing elective surgical treatment of lung cancer were randomly assigned to two group. In the PCEA group, 1.2M lidocaine was infused to epidural space at a rate of 5 ml/h and a bolus of 2 ml/15 min interval time. In the PCIA group, fentanyl was infused intravenously at lock rate of 0.3μg/(kg-h) (lm l), and a bolus of 1 ml/15 min interval. The degree of pain control was evaluated at four levels, i.e. grade 1 with effective expectoration and no pain; grade 2 with effective expectoration and light pain;grade 3 with rescue drug to relief the pain during effective expectoration and grand 4 with continuous pain as breath and sleep was interfered. Additional treatment sedation and analgesia was needed. Physical sign and unwanted side effect were also recorded. Results : In the PCEA group patients with grade 1 or 2 accounted for 86.7% (20/30), and patients with grade 3 (3) and grade 4 (prolapse of epidural tube was not recorded). The incidence of nausea and vomiting was 10% (3/30) in PCEA. In the PCIA group pain-relief patients with grade 1 or 2 accounted for 36.6% (11/30) only. The incidence of dizziness, nausea and vomitting was 76.6% (23/30) and the other sign was normal. Conclusion : The effect of PCEA is better than that of PCIA in pain control after surgical treatment of lung cancer.