Abstract: In order to solve the problem of seismic damage prediction of underground pipeline system,which lacks investigation data in planning urban hazardous prevention,rough set methodology was adopted to analyze the problem.Firstly,ROUSTIDA algorithm was used to complete missing values in decision table of underground pipelines with incomplete information system.Secondly,an improved Greedy algorithm was made use of the discretization of decision table and influenced factors reduction algorithm based on condition information entropy was introduced to reduce attributes and obtain information to build rule base, and then seismic damage prediction of underground pipeline system can be analyzed according to the rule database． Finally, taking the underground pipelines in Quanzhou as an example, after doing lots of prediction experiments and comparing with other common prediction methods, such as the theory methodthe test result denoted that the recognition accuracy can reach 90％,and the method proposed in the paper proved to be effective and feasible．