There is a lot of redundant information in a data cube. Removing redundancy from a data cube can not only reduce the storage space but also accelerate the computation. Tuples of a data cube can be divided into closed-tuples and non-closed tuples. For any non-closed tuple, there exists a closed-tuple, and both are aggregated from the same set of tuples in a base table and have the same aggregated value. By removing all non-closed tuples, a data cube can be translated to a closed data cube. The algorithm of computing a closed data cube is given, answering a query and maintaining the closed data cube incrementally. The results of experiments are also presented by using both the synthetic and real-world data sets. The experimental results show that the closed data cube technique is effective.