Abstract: In order to improve the robustness of network attacks in strong uncertainty environment, a robust strategy against node attacks based on seeding perception is proposed. Three diffusion entropy measures to evaluate the diffusion robustness of probabilistic networks were introduced to detect the strength of attack robustness under independent cascade or linear threshold model. The concept of seeding perception was introduced, and an algorithm to measure the network robustness under different perception levels was proposed, so as to calculate the impact of node attacks and improve the computational efficiency. The experimental results show that the proposed method can effectively evaluate and enhance the robustness of probabilistic networks.