To evaluate the effect of hepatic resection for metachronous liver metastases after resection of gastric cancer, the clinicopathological factors of gastric cancer, state of liver metastasis, surgical procedures for liver metastasis, and remote survival were studied. Between 1989 and 2001, 30 consecutive patients underwent hepatic resections (36 resections) for metachronous liver metastases. The patients included 25 men and 5 women, and the median age was 60 years old (range 46-86 years old). As for curability, curative A/curative B was 10/20, and H0/H1/H2 was 25/3/2. The mean period from initial surgery to the liver resection was 19.3 months (range 6.3-65.2 months), and the liver metastatic conditions were H1 for 27 patients and H2 for 9 patients. By number of liver metastases, 27 patients had 1 lesion, 24 patients had 2, and 5 patients had more than 3 lesions. Twenty-nine patients were negative and 7 were positive for lymph node metastasis. A partial resection was performed for 13 patients, a subsegmentectomy for 3 patients, a segmentectomy for 7 patients, a lobectomy for 8 patients, and an extended lobectomy for 5 patients. For all patients except those who had other causes of death, the overall mean survival time was 702 days and the 5-year survival rate was 26.4%. There were four 5-year survivors after hepatic resection. In conclusion, the main prognostic factor after resection of metachronous liver metastases was the existence of lymph node metastasis, and the remote survival of patients with abdominal para-aortic lymph node recurrence was poor. The type of liver resection was not a predictor of survival. The resection of liver recurrence was evaluated clinically, but we should maintain strict criteria and select adequate surgical procedures.