Objective To explore a safe and effective way to establish reflux esophagitis model consistent with the anatomical structure. Methods Forty-five New Zealand rabbits aged 5 months were randomly divided into three groups. The rabbits undergoing lower esophageal sphincter incision were assigned to A group, those undergoing partial cardiomyectomy to B group, and those undergoing balloon dilatation by Forley catheter to C group. All the rabbits underwent esophageal fistulation before establishment of reflux esophagitis model. 24 hour-pH monitoring was performed to all the rabbits before and at 2 weeks postoperatively. At 4 weeks postoperatively, the animals were sacrificed for histological and electron microscopic results. Results The mortality rate in B group was higher than those of the other 2 groups (P=0.057). With respect to the total time ratio of pH <4 (%) in 24 hours, number of reflux episodes and number of reflux episodes >5 min, a clear trend was seen as B group > C group > A group; however, there were no significant differences among the 3 groups. The success rates of establishing reflux esophagitis model were similar among the 3 groups. For the reflux esophagitis models, moderate/severe esophagitis was more commonly-seen in B group than in the other 2 groups, but showing no significant difference. Conclusion Partial cardiomyectomy is a more effective way to establish a reflux esophagitis model, but with higher mortality. Balloon dilatation is a safe and effective to establish a reflux esophagitis model, which is also more consistent with anatomical structure and suitable for research of endoscopic or other new surgeries.