Objective: Our approach is to reduce the large data size to a small data size which represents same features of the total large data set, so that computational complexity becomes shorter. Method: In this paper we present a new approach to minimize the data size and then to cluster that reduced data. The volume of data being generated nowadays to cluster is increasingly large. How to extract useful information from such data collections is an important issue. A promising technique is the Rough set theory, a new mathematical approach to data analysis based on objects of interest into similarity classes which are indiscernible with respect to some features. Result and conclusion: This theory offers two fundamental concepts: reduct and core. In this paper, some basic ideas of rough set theory are first presented. Some experiment results are also given.