Objective To compare the effects of different dietary intervention regimens and the combined effect of metformin after the establishment of a type 2 diabetes mellitus (T2DM) model, to establish and maintain a T2DM rat model that is more consistent with clinical practice. Methods Fifty male SD rats were divided randomly into control (n= 10) and model groups ( n= 40). Rats in the model group were fed a high-sugar and high-fat (HSHF) diet, plus daily intragastric administration of fat emulsion starting from the 3rd week. After 4 weeks, T2DM was induced by two intraperitoneal injections of streptozotocin 25 mg / kg. After successful model establishment, rats in the model group were divided randomly into four subgroups ( n= 10 rats per group): HSHF model ( HSHF-M),normal diet model (ND-M), HSHF metformin (HSHF-Met), and normal diet metformin groups (ND-Met). Rats in the drug-treated groups received dietary intervention combined with metformin, while rats in the model groups received dietary intervention alone, and rats in the control group were fed a normal diet. All interventions lasted for 12 weeks. The general status of the rats and their survival indicators were monitored during the dietary-intervention period. After 12 weeks, an oral glucose tolerance test (OGTT) was performed. Serum fasting insulin ( FINS) was detected to calculate the insulin resistance index (HOMA-IR) and islet β-cell function index (HOMA-β). Serum levels of blood lipids, liver function, renal function, oxidative stress indicators, and inflammatory factors were also measured and pathological changes in the pancreas, liver, and kidney tissues were observed and analyzed. Results After successful establishment of a T2DM rat model, feeding with a normal diet maintained model stability within 12 weeks. Compared with the HSHF-M group, rats in the ND-M group had significantly increased body mass (P<0. 05) and significantly decreased urine output and blood glucose (P<0. 05). Regarding the OGTT, blood glucose levels at 0,30, 60, and 120 min were significantly lower in the ND-M compared with the HSHF-M group and the area under the curve was significantly reduced ( P<0. 05). Regarding blood glucose-related indicators, there were significant differences in glycated serum protein, FINS, and HOMA-IR (P<0. 05). For blood lipid-related indicators, total cholesterol, triglycerides, free fatty acids, and low-density lipoprotein cholesterol levels were all significantly decreased (P<0. 05), while high-density lipoprotein cholesterol was significantly increased (P<0. 05) in the NDM group. Liver function-related indicators (alanine transaminase, alkaline phosphatase, aspartate transaminase) and renal function-related indicators (creatinine, blood urea nitrogen) were significantly decreased (P<0. 05), while nitric oxide levels were significantly increased (P<0. 05) in the ND-M group. Regarding oxidative stress indicators, malondialdehyde levels were significantly decreased (P<0. 05) and superoxide dismutase was significantly increased (P<0. 05) in the ND-M group. Levels of inflammatory factors (tumor necrosis factor-α, IL-1β) were significantly decreased (P<0. 05) in the ND-M group. Histopathological changes, such as hepatic cord disorder, islet cell reduction, and glomerular vacuolation, were significantly alleviated in the ND-M group, in the absence of long-term intake of an HSHF diet, compared with the HSHF-M group. Survival and biochemical indicators were significantly improved in the HSHF-Met and ND-Met groups compared with the HSHF-M group in ( P<0. 05), with a more significant improvement in the ND-Met group while pathological damage to the pancreas and liver was less severe than in the HSHF-Met group. Conclusions Feeding a normal diet after establishment of a T2DM model can effectively avoid glucose and lipid metabolism disorders, oxidative stress, inflammatory responses, and multi-organ pathological damage caused by a HSHF, thereby allowing the successful establishment and maintenance of a T2DM rat model that is more consistent with clinical pathways.