One hundred and twenty women undergoing gynaecological abdominal operations were randomized to receive either epidural bupivacaine 0.0625% + fentanyl 3.3 micrograms/ml infusion (Group EPI, n = 57), or patient-controlled intravenous morphine analgesia (Group PCA, n = 54) for postoperative pain relief. The groups were comparable in demographic data, types and duration of operation. Group EPI achieved significantly lower verbal rating scale of pain (VRS) at rest at 0, 4, 12, 16, 20, 28 and 40th postoperative hours. The VRS during cough were also significantly lower in Group EPI at 0, 4, 8, 12, 28 and 36th postoperative hours. None of the patients had respiratory depression or hypotension. Nausea/vomiting occurred in 52.6%/33.3% of patients in Group EPI and 52.7%/37.0% in Group PCA. Most patients (84.2% in Group EPI and 72.2% in Group PCA) rated their pain management as "good". We conclude that epidural infusion of bupivacaine 0.0625% and fentanyl 3.3 micrograms/ml provide better analgesia than patient-controlled intravenous morphine after gynaecological laparotomy.