In recent years, China has experienced frequent chemical production accidents. This study collates 1900 briefings of such accidents from 2012 to 2023, sourced from a variety of repositories. By employing association rule mining, we analyzed the connections between causative factors and patterns of these accidents. The analysis revealed significant association rules characterized by high lift values, severe consequences, and patterns not previously recognized. A network model was constructed utilizing Gephi® software to represent the causative factors of these accidents. Through a centrality analysis of the network nodes, key factors contributing to these incidents were identified. Moreover, a SARIMAX model was developed and validated using time series data to predict future accident trends in chemical production. The forecasts generated by this model provide valuable insights for chemical production sectors, highlighting periods with an increased likelihood of accidents. Conclusively, this integration of data mining and predictive modeling could provide a critical method for improving safety protocols and enhancing risk management in chemical industry.