Objective To investigate the mechanism of gastric bypass surgery in the treatment of type 2 diabetes mellitus in a rat model. Methods Seventy-two 8-week-old GK rats were randomly divided into operation group, sham operation group, diet control group and control group (18 rats in each group) according to the random number table. Rats in the operation group and the sham operation group received gastric bypass surgery and transection and reanastomosis of the gastrointestinal tract, respectively. The food intake was set as 15 g/d for each rat in the diet control group, while rats in the control group were fed ad libitum. The levels of fasting blood glucose ( FBG), postprandial blood glucose (PPBG) and glucagon-like peptide-1 (GLP-1) were detected before operation and at postoperative week 2, 4 and 8. The levels of PPBG and GLP-1 were detected at postoperative week 2, 4 and 8, then 6 rats of each group were sacrificed to detect the apoptosis of islet B cells using the TUNEL method. All data were analyzed using the t test. Results In the operation group, the preoperative levels of FBG and PPBG were (16.2±0.8)mmol/L and (31.1 ± 1. L)mmol/L, respectively, which were significantly higher than (9.2± 0.6) mmol/L and (13.1 ±0.7) mmol/L at 4 weeks after the operation, and (9. 7 ± 0. 7) mmol/L and (12. 3 ± 0.7) mmol/L at 8 weeks after the operation (t = 20. 7, 49. 7; 18. 8, 39. 0, P < 0.05 ). The levels of FBG and PPBG before the operation and at 4 and 8 weeks after the operation in the operation group were significantly lower than those in the sham operation group, diet control group and control group at corresponding time points (t = 27.7, -57.8; 11.3, -59.9; -27.4, -48.2; -13.2, -52.7; -7.0, -24.9; -18.2, -56.4, P<0.05). In the operation group, the levels of fasting GLP-1 and postprandial GLP-1 were ( 10. 7 ± 1. 0) pmol/L and (42.5 ±1.2)pmol/L, respectively, which were significantly lower than (26. 1 ±0.9)pmol/L and (90.7 ± 1.7)pmol/L at4 weeks after the operation, and (25.3 ± 1.2)pmol/L and (90.4 ±2.0)pmol/L at 8 weeks after the operation (t=42.1, -92.4; -29.1, -72.7, P <0.05). The levels of fasting GLP-1 and postprandial GLP-1 before the operation and at 4 and 8 weeks after the peration in the operation group were significantly higher than those in the sham operation group, diet control group and control group at corresponding time points (t = 48.0, 61.9; 38.0, 62.2; 50.9, 65.2; 37.0, 48. 1; 27.5, 51.6; 17.5, 52.9, P<0.05). The number of the apoptotic islet β cells in the operation group was decreased with time. The apoptosis rates in the operation group, sham operation group, diet control group and control group were 5.9%±0.7% , 47.2%± 1.0% , 21. 1%± 1. 2% , 46.5%±1.4% at 4 weeks after the operation, and 6.3%±1. 1% , 47.2%±1.0% , 21.2%±1.2% and 46.0% ± 1.4% at 8 weeks after the operation. The apoptosis rates in the operation group were significantly lower than those in the sham operation group, diet control group and control group at corresponding time points (t = -82. 2, - 67. 0; - 27. 1, - 22. 4; - 55. 2, - 54. 6, P < 0.05). Conclusion After gastric bypass surgery, the level of blood glucose reduces and the level of GLP-1 increases which significantly inhibit the apoptosis of islet B cells in rats with type 2 diabetes mellitus.