We compared the outcome of two surgical alimentary canal reconstruction methods after proximal gastrectomy. Three hundred ninety-six patients who underwent a radical proximal gastrectomy were randomized into two groups. Group A was treated with modified double tracks anastomosis, and Group B was treated with esophagus-remnant stomach direct anastomosis. Outcome measures were hematological indices, prognostic nutritional index (PNI), plasma hormone concentrations, and Visick index. The operative times in Groups A and B were 210 +/- 53 and 150 +/- 75 minutes and the hemorrhage volume was 173 +/- 67 and 150 +/- 75 mL, respectively. Six months after operation, values of hemoglobin concentration, body weights, and PNI indices were significantly increased in Group A compared with Group B. Levels of gastrin and somatostatin were obviously less than preoperative values and levels of cholecystokinin and motilin were significantly higher than preoperative values in both groups. All patients of Group A were classified into Visick index Grades I and II, whereas only 70.37 per cent of Group B belonged to Visick index Grades I and II postoperatively. The overall 3-year survival rate was not significantly different in both groups. Modified double tracks anastomosis as an alimentary canal reconstructive method for radical proximal gastrectomy showed better outcomes than esophagus-remnant stomach direct anastomosis.