Abstract: In working environment of greenhouse with high temperature and humidity，the sensors are prone to failure，which causes a certain waste of resources for intelligent management of greenhouse.To ensure effective operation of intelligent greenhouse management system，it is necessary to identify faulty sensors in greenhouse in a timely and accurate manner.Through experiments，changes in the indoor field of Yunnan greenhouses under different ventilation conditions in summer were analyzed.Taking large glass multi-span greenhouses in Yunnan as research object，a three-dimensional steady-state CFD model of greenhouses under different ventilation modes in summer was established and combined with experiments.Validity of the model was verified by experiments.According to actual external conditions，internal field of greenhouse was analyzed to predict environmental value at sensor node position，and abnormal position of sensor reading was found by comparing with actual reading.In order to further improve accuracy of fault sensor identification，combined with sensor fault node detection algorithm based on median strategy，it could accurately identify greenhouse fault sensor in time and provide accurate greenhouse environmental data for intelligent management of greenhouse.