Because of the strong time dependence and spatial correlation of the data collected by the greenhouse wireless sensor networks (WSN), according to the basic structure of the greenhouse WSN, the mathematic model of fault diagnosis based on the time series and space series analysis was set up respectively through preprocess sample information of the time series and space series on the basis of analyzing the raw data of WSN. According to this, the working condition of the sensor node was analyzed and the fault diagnosis algorithm of the greenhouse WSN was given. The research result showed that the method of fault diagnosis could find out the anomaly of greenhouse WSN and judge the fault node, thus enhance the reliability of WSN. 