Wireless sensor networks（WSNs） generally adopts battery as power

research of WSNs focus on energysaving. To improve precision of measurement result and reduce network energy consumption

a two-layer mode data fusion program is proposed. Low-level data fusion based on Grubbs criterion and the order of weighted algorithm is used on sensor nodes

high-level data fusion based on neural network algorithm is used on cluster head nodes. The simulation experimental results show that the two-layer mode data fusion scheme effectively reduces amount of data transmission

improves precision of the fusion data and reduces energy consumption of sensor nodes.