Knowledge reorganization is one of the key links in supporting decision-makers to acquire the information of
decision-making in risk management. This process provides decision makers with textual knowledge which has strong
correlation with decision-making problem situations. This paper proposes a knowledge reorganization method based on
categorical structure semantic correlations based on genetic algorithm. This method provides a new way for the highdimensional
semantic information processing which are necessary for the quick knowledge acquisition of risk decisionmaking.
The process of genetic evolution is achieved through four aspects including genetic coding, operators, convergence
control and parameter selection. Simulation experiment results show that this method can obtain better reliability of
reorganization under forced convergence conditions.