For the large number of sensors are easily influenced and even broken by environment interference

new advanced fusion technique is requested for WSNs.Median filter is introduced

its nicer ability to restrain pulse noise is made the use of

Kalman filter is combined with to research new fusion algorithm adapted to WSNs.Take spatial-temporal fusion to reduce the cost of computing

which makes the network have real time ability.The algorithm can process the data with mistakes

so that it can improve the robot of the network.The validity of the algorithm is shown by simulation.