Abstract:As the sparse deployment of sensor nodes in the project, the senor data don’t obey statistical law. The traditional method of removing negligent errors only processing data from the view of mathematics, can't reflect the characteristics of data changes with time. And this kinds of methods require the data to normal distribution which unable to effectively deal with these reality application data. This paper proposes taking amplitude-limiting filtering algorithm applied to WSN data pretreatment, to solve the above questions. The experiments verify that the use of amplitude-limiting filtering algorithm WSN data pretreatment has better leverage error rate, the rate of false positives search even lower.