Aiming at that the safe retrieval structure of cloud data query fails to consider the uncertainty of data, one can not stably mine data features in the interference of massive data redundancy. This paper proposed a method for security index construction of cloud data query specializing in the interference of massive data redundancy. The construction rule is based on the data security query. This method could establish a new safe index structure of cloud data query, eliminate the redundant data through the iterative method, then index individual cloud data from R tree in the central region in proper sequence, so as to get more precise and more secure query or index results. Experimental results show that the cloud data index scale of proposed method is large, and data query efficiency and precision ratio are high, which has good antiinterference performance. This method can eliminate massive interference of redundant data in the process of index, which has good stability.