Abstract: The possibility of occurring exception is relatively small in most applications of wireless sensor networks. So data obtained in sequent moment by the same node have time correlation, and data obtained in the same time by adjacent nodes have space correlation. A large number of energy of node will be wasted if data which include time and space correlation is transmitted. Therefore, this paper proposed a data compression algorithm for wireless sensor networks based on optimal order estimation and distributed clustering. The algorithm explores the time and space correlation among data obtained by sensors. The correlation parameter can be get based on optimal order estimation. Then all data can be restored based on time and space correlation parameters and only a little necessary data are transmitted by nodes. Because redundancy is decreased when data is transmitted, the average energy cost of node is reduced and the life of the whole wireless sensor networks can be extended.