In view of the existing problems in traditional greenhouse, such as difficulties in remote operations, necessity of manual intervention in data collections, and low intelligence in production, an intelligent edge Mesh sensor network was constructed based on edge computing, and a cluster head selection method for temperature and humidity sensor network in greenhouse was proposed based on improved LEACH algorithm. The temperature and humidity sensor nodes were constructed by using ESP8266-12F wireless module, NodeMCU type based Internet of Things extension board and DHT-11 temperature and humidity sensor, and the automated data acquisition algorithm was developed. Based on the ESP8266-12F wireless module, the edge wireless Mesh sensor network was constructed, and the automatic networking between nodes was achieved. Aiming at the problems of large load consumption and low signal transmission rate of central router, a method of dividing the network based on RSSI value of central router was proposed, which improved the transmission rate of the network. A temperature and humidity sensor network Sink cluster head selection algorithm was proposed based on LEACH algorithm, which was suitable for the plant factory temperature and humidity sensor network deployment and beneficial to balance the overall energy efficiency of the network. The simulation experiment results showed that when the probability of the initial cluster head was set to be 0.1, by using the original LEACH algorithm, the probability was 10.86% for the cluster head appearing in the center of the sensor network; however, it was increased to 17.42% when using the improved LEACH algorithm with weight k=1; and the probability would be further increased to 24.96% for the cluster head appearing in the center position when using the improved LEACH algorithm with weight k=2.