Abnormal or drastic changes in the natural environment may lead to unexpected events, such as tsunamis and earthquakes, which are becoming a major threat to national economy. Currently, no effective assessment approach can deduce a situation and determine the optimal response strategy when a natural disaster occurs. In this study, we propose a social evolution modeling approach and construct a deduction model for self-playing, self-learning, and self-upgrading on the basis of the idea of parallel data and reinforcement learning. The proposed approach can evaluate the impact of an event, deduce the situation, and provide optimal strategies for decision-making. Taking the breakage of a submarine cable caused by earthquake as an example, we find that the proposed modeling approach can obtain a higher reward compared with other existing methods.