Abstract: To explore new ways of getting security situational factors for complex large-scale networks,a danger theory based model for network attacks situation perception with cloud method is proposed.The proposed model follows the danger model principles of cells dying unnaturally or distressed may release an alarm signal,and the adaptive immune response stimulated by danger signal.With the establishment of mapping the computer networks to immune danger model,the model is built by using the uncertainty reasoning method of cloud computing.Within the model,the quantitative safety state parameters,such as resource occupancy rate,network traffic are translated into different qualitative risk levels by cloud generators.Network attacks’ detection is activated by danger signal,which is generated by different risk level.Theoretical analysis and simulation results show that the presented model is valid.Thus,it provides a good solution to large-scale computer network security situation awareness.