Abstract: The paper focused on the risk identification of small probability events．A Residual Antigen based Dynamic Memory Risk Identification Model DMRIM was proposed．As small probability events have no rules to follow,the intensity and frequency of risk is intuitively and dynamically mapped to the concentration of residual antigen by DMRIM． By using residual antigent,DMRIM stimulates immune memory,guides antibody evolution and controls the life-cycle of the detector．It breaks through traditional memory cell life-cycle and realizes distribution and self-made of the detector,thereby improving recognition of small probability events．The simulation experiments prove that DMRIM achieves dynamic memory and recognize small probability events effectively．Its feasibility is verified in practical application.

Key words: Residual antigen,Imunne memory,Dynamic risk identification,Small probability