Safety production is a key concern of mining enterprises. In order to manage production more effectively and ensure the safety of underground operation, many mines have established safety big data platform. However, due to the lack of effective analysis model, many platforms only store data and do not use hidden danger data effectively. Therefore, Therefore, based on big data analysis technology, this paper constructs a multi-dimensional analysis model of mine safety hidden danger, and analyzes the distribution law of hidden danger in time and space; classifies many hidden danger information by using topic mining model, and obtains 15 hidden danger topics; explores the internal relationship between different hidden dangers by using association rule mining model, and the R programming language is used to visualize the above results. The analysis and study of the hidden danger not only makes full use of the hidden danger data, avoids the waste of data resources, but also has certain guiding value for the prevention of mine accidents.