Background To improve the therapeutic effects of gastric electrical stimulation (GES) for obesity, an animal experiment was conducted using a new type of stimulators. Proper parameters of GES were selected, and the impacts of GES on the food intake and gastric accommodation of canines were observed. Methods Eight beagle dogs were operated on, and GES was performed on them. Firstly, GES was performed to determine the right parameters according to symptoms. Secondly, the so selected parameters were used in a 3-day GES procedure, during which process food intake, body weight, and symptoms were recorded. Thirdly, the gastric capacities before and after GES with different pulse widths were measured by means of a barostat. Results The selected parameters varied for each dog, with the pulse widths ranging from 0.3 to 6 ms. The food consumption after GES dropped significantly as compared with the amount observed in the sham stimulation. Tolerance to stimulation could be observed during GES. The post-GES gastric fundus capacity increased evidently in comparison with the capacity before GES, suggesting significant distention as compared with sham stimulation. Given an increment of 2 ms in the pulse width twice, the gastric capacity continued to distend each time. Conclusions GES featuring pulse trains with wider and individualized pulse widths could inhibit food consumption of dogs. The stimulation parameters should be selected individually and adjusted periodically. GES of this mode could also increase the fasting gastric capacity with certain dose-related effects. The new type of stimulators may be more suitable for the treatment of human obesity than traditional stimulators.