Abstract: Wireless sensor networks have been widely utilized in agricultural production in such as crop information monitoring systems, agricultural facilities’ wireless control systems, etc. The wireless sensor networks could promote the development of agricultural information and intelligence, and more research has been focused on using ZigBee wireless technology to build the networks in recent years. To collect the base crop status and environmental information of greenhouses in a wireless way, a wireless sensor monitoring network system was designed. The basic work of this paper was the software and hardware system design; further work is projected to be low-power adaptive mechanism design. In analyzing the distribution characteristics of the greenhouse base, it could be seen that the greenhouses were concentrated in their distribution, but independent from each other. So the network topology architecture was designed as clustering Tree-Mesh hybrid topology architecture, and the nodes of the cluster belonged to the same greenhouse. The network was built up by a coordinator, and a large number of routers and sensor nodes were joined in. The coordinator was a sink node, it was designed as a gateway, and there were some routers which played the role of cluster head in the network. The clustering Tree-Mesh hybrid network was built in two steps: First, the mesh network was established by the coordinator and cluster head. Then, the tree network was built by the cluster head, and the tree was a cluster with routers and sensor nodes. The system utilized ZigBee to build the wireless sensor network and multi-hop communication, and the hardware of a single chip multi-sensor wireless node based CC2530 was designed. The modular design of the hardware subsystem was composed of a radio module, sensor module and power module. The finite state machine node software and the low-power improvement were designed based on Z-Stack. The stack ran on a task allocation mechanism that was similar to the embedded operating system. According to the wireless channel quality of the motionless node, the best transmit power adaptive mechanism based Received Signal Strength Indication was designed, and the node adjusted the transmission power to get a credible communication link. According to the energy consumption distribution of the node subsystem, to reduce the energy consumption of the wireless transmission, a minimum data transmission redundancy adaptive mechanism based perceptual data difference value was designed. And the node dropped the sensor data which had less information to avoid frequent wireless transmission. With these adaptive mechanisms, the node could meet the requirements of low power consumption and low network data redundancy. The test results showed that, when the transmitting power is 1 mW, the typical data rate of point-to-point communication is 20 kb/s, and of multi-hop is 0.3 kb/s. The communication distance of the node with the 5th battery powered is up to 30 m, and DC is 90 m. The theoretical simulation analysis showed that the energy consumption of the low-power adaptive mechanism node reduces by 38.44%. The system can meet the greenhouse base environmental monitoring application.