Abstract: A prediction method is proposed to cope with abnormal event numbers, which often appear in chemical process due to external disturbance. If an abnormal event is not effectively controlled，it will probably result in an accident. The higher abnormal event numbers are, the greater the probability of production accidents is. Therefore, the precise prediction on the abnormal event numbers can effectively improve the risk management level on the chemical process. Usually, four operating teams work in a workshop and the number of abnormal event varies from team to team. Based on operating teams, the dynamic prediction model is constructed by using the Bayesian theory and Vine Copula such that the abnormal event numbers in a operating team can be predicted.