A dynamic Risk Management (RM) provides monitoring, recognition, assessment, and follow-up action to reduce the risk whenever it rises. The main problem with dynamic RM (when applied to plan for, how the unknown risk in unexpected conditions should be addressed in information systems) is to design an especial control to recover/avoid of risks/attacks that is proposed in this research. The methodology, called Dynamic Intelligent RM (DIRM) is comprised of four phases which are interactively linked; (1) Aggregation of data and information (2) Risk identification (3) RM using an optional control and (4) RM using an especial control. This study, therefore, investigated the use of artificial neural networks to improve risk identification via adaptive neural fuzzy interface systems and control specification using learning vector quantization. Further experimental investigations are needed to estimate the results of DIRM toward unexpected conditions in the real environment.