Abstract: To improve the dynamic and precise risk control in coal mining faces, this study proposes a dynamic risk classification and control model based on Case-Based Reasoning (CBR), addressing the limitations of traditional risk assessment methods that struggle to adapt to real-time changes in underground environments and neglect the reuse of historical experience. By constructing a risk case base covering four-dimensional characteristic attributes (personnel, equipment, environment, and management), quantifying indicator weights using the grey clustering method, and designing a case retrieval and similarity calculation algorithm based on text matching, the model achieves rapid matching of new risk cases and reuse of control measures. Taking the 8511 working face of Chenjiagou Coal Mine as an example, the model's application in roof and gas risk scenarios was tested: case retrieval similarity reached 100%, allowing precise reuse of support reinforcement, gas monitoring, and other control measures, with the accuracy of risk level determination increasing by 23%. The research demonstrates that the model significantly enhances the timeliness and specificity of risk response through dynamic case base updates and multi-attribute matching mechanisms, providing intelligent decision-making support for coal mine safety production.