In order to improve the current data mining method of the dam safety monitoring database which runs slowly and takes up a lot of computational space, a modified FP-growth algorithm was proposed. The pre-processed monitoring data was pruned, and then frequent item mining was performed after the Priority tree was generated. In the application of exploring the correlation between dam deformation and water temperature and other environmental quantities, the proposed method not only has high mining speed, high precision, and simple results, but also can compare a single factor or analyze the relationship between multiple factor coupling with target variables. The example shows that the improved FP-growth algorithm provides a good idea for dam safety monitoring data mining.