Abstract: Continued developments in Chinese economy and society have led to a rapidly changing meat supply-demand system．Accurate and long-term prediction of meat supply-demand has become a concern for scholars．Quantitative research on the supply and demand of meat in China involves two main model types: time series-based models and evolutionary attribution models．Meat supply-demand is complex, previous works often used simplifications, unreflective of the complex nature of meat supply-demand．Existing supply-demand models on meat supply and demand indicate that differences among regions must not be ignored, machine learning is reliable in the meat supply-demand field．Meat consumption, production, trade and government policy are dynamically balanced at all times．Regional differences and historical research are taken into account, forecasts should be carried out on consumption, production, trade and policy modules respectively．Balanced interactions between modules should be simulated using multi-agent thinking．