Objective To evaluate the effectiveness of intravenous patient-controlled analgesia versus epidural patient-controlled analgesia for postoperative analgesia, sedation, and the incidence of side effects. Methods We searched the specialized trials registered in the Cochrane anesthesia group, The Cochrane Library (CCTR), MEDLINE (1966 to Sept. 2008), EMbase (1966 to Sept. 2008), PubMed (1966 to Sept. 2008), and handsearched some Chinese anesthesia Journals and Clinical anesthesia journals. Randomized controlled trials (RCTs) and quasi-RCTs of intravenous versus epidural analgesia for post-operation were included. The quality of the trials was critically assessed. RevMan 4.2.8 software was used for meta-analyses. Results Thirteen RCTs involving 580 patients of intravenous versus epidural analgesia for post-operation were included. The results of meta-analyses showed that there were no significant differences in postoperative analgesia and sedation at the hour-points of 2nd, 4th, 8th, 12th, and 24th hours after operation. There were no significant differences in plasma fentanyl concentration in the two groups on the same analgesia effects. There were also no significant differences in side effects. Conclusion Both intravenous patient-controlled analgesia and epidural patient-controlled analgesia have the same clinical effects. Compared with epidural patient-controlled analgesia, intravenous patient-controlled analgesia has fewer side effects and is more convenient. At the same time, it can reduce more costs of hospitalization. But because of the low quality and small sample size of the included studies, more well-designed, large scale, randomized controlled trials are needed.