Abstract: Yunnan Province, as the core producing region of China's coffee industry, dominates the national market in terms of coffee production. In recent years, significant fluctuations in Yunnan's coffee production have had a certain impact on the entire industrial chain. In view of this, it is particularly important to conduct an early warning analysis of coffee supply to ensure the stability and security of Yunnan's coffee industry. This study aims to forecast and risk assess Yunnan's coffee supply by constructing an effective early warning model. The study first determines the annual growth rate of Yunnan coffee production as the warning indicator, and selects the key factors affecting Yunnan coffee supply as the warning indicators. In order to further analyze the dynamic relationship between these indicators, the study uses hierarchical cluster analysis to classify the indicators into leading, synchronous and lagging indicators. Subsequently, the reasonableness of the selected indicators was assessed using grey correlation analysis to ensure the scientificity and practicality of the model. On this basis, this study constructed the ARMA (6,2) early warning model. The model was trained and validated with historical data from 2006 to 2021. After the model training was completed, we used the model to predict the growth rate of Yunnan coffee production from 2022 to 2030. The prediction results show that there is a high degree of agreement between the predicted values of the ARMA (6,2) early warning model and the actual observed values, which indicates that the constructed model has high accuracy and application potential in Yunnan coffee supply early warning.