Prior knowledge serves as some relational patterns assisting to recognize connections between apparently independent events and trends, which is the main way of prior knowledge working in opportunity discovery process. However, these relational patterns usually are latent, imprecise and semi-structured, are difficult to be formal description. This results in restriction of current methods in Artificial Intelligence to support opportunity discovery. To solve this problem, a hypergraph model is proposed to describe and construct the relational patterns, within which discrete vertices and latent relations are obtained by text association mining. A case of China’s commercial bank’s restructure is used to describe the application of it. The result shows that the model has abilities of visualization and simulation for the components and patterns mined from texts, as well as supporting the opportunity discovery. The further research of this model is mentioned in the end of paper.