Abstract: To reduce the energy consumption and extend the network’s survival time is the key research problem in the Wireless Sensor Network（WSN）.Much research work has been done on the placement of nodes and routing protocal，reducing data redundancy by density controlling and hierarchical clustering algorithm.In this paper，it is proposed that multi-sensor information fusion is an effective way to reduce the communication data and save the energy consumption.Three distributed sensor fusion methods are researched：fusion with weighted coefficients，Kalman fusion filtering without feedback and Kalman fusion filtering with feedback.The three algorithm are implemented in a temperature-measurement WSN.And then the three fusion methods are compared and analysed.Finally，the effectivity and conditionality of the fusion methods are summarized.