BACKGROUND: Approximately 537 million adults worldwide have diabetes, more than 90 % of which is type 2 diabetes mellitus (T2DM). China has the largest number of people living with diabetes. Understanding the epidemiological mechanism can guide diabetes surveillance and control. METHODS: Utilizing the most recent Global Burden of Disease 2021 (GBD2021) data on T2DM and risk factor exposure values, we developed a modelling framework that employs systems of difference equations to project the future burden of diabetes among the Chinese population. The model characterized the diabetes progression process from no diabetes, undiagnosed diabetes, and diagnosed diabetes with and without complications, in which genetic and lifestyle factors modulate the rate of development at these stages. We focused on the long-term dynamics of diabetes progression with the impacts of influence factors. We then fit the model to the longitudinal numbers of T2DM patients by Markov chain Monte Carlo (MCMC) algorithm. RESULTS: The model with the influencing factors fitted an R CONCLUSIONS: This study provided a modelling framework to explore the mechanism of T2DM development, which allows to quantify the impacts of risk factors on T2DM progression. The results highlight the high burden of T2DM in China and emphasized the importance of lifestyle interventions.