A randomized controlled trial of hepatic arterial embolization was conducted in 63 consecutive patients who had unresectable but still embolizable hepatocellular carcinoma. Patients were randomized into three groups. Patients in group 1 received multiple hepatic arterial embolizations; patients in group 2 were given hepatic arterial embolization once, followed by monthly chemotherapy with high doses of 5-fluorouracil; and patients in group 3 received only monthly chemotherapy with high doses of 5-fluorouracil. Complete response was achieved in only 1 patient who received multiple hepatic arterial embolizations. Partial responses were observed in 13 patients (61.9%) in group 1, 10 patients (47.6%) in group 2, and 2 patients (9.5%) in group 3. The survival rates of patients in group 1 at the end of the ninth, 12th, 15th, 18th, and 21st months were 53.2%, 42.2%, 42.2%, 42.2%, and 42.2%, respectively, which were not significantly different from those of patients in group 2 but were better than the survival rates of patients in group 3. The results suggest that hepatic arterial embolization is an effective palliative treatment that prolongs survival of patients with unresectable hepatocellular carcinoma.