Abstract: The low-power wide area networks (LPWAN) provide efficient and feasible technical means for monitoring the environment of island aquaculture. At present, most successful environment monitoring applications have been deployed in artificially controlled environments such as cities and parks, or with the help of mobile telecommunication networks. Unlike previous researches, this study introduces LPWAN into wild environments such as islands and integrates with existing sensor technologies to build long-range wireless sensing and monitoring systems. Due to the relatively scattered areas, adverse working environments, inconvenience of manual inspection and long period of data collection, several problems exist if traditional wireless sensor network (WSN) is used for the deployment of island aquaculture environmental monitoring system. Firstly, the communication range of the node is limited; secondly, the network needs too many routing nodes and is too complex; thirdly, the overall network power consumption will be rather high. In order to improve the performance of the system, the paper designed a new kind of monitoring system for island aquaculture environment based on LPWAN. The system includes terminal acquisition nodes with integrated Arduino and some sensors, a sink gateway for data aggregations and long-distance transmission via the long range radio (LoRa), and a background database management system for data collection, storage, processing, access and control using Python and PostgreSQL. When the monitoring network is actually deployed, it is considered that some nodes locate in the center of the aquaculture area and the monitoring worker cannot easily reach the locations. Therefore it is difficult to easily place and deploy the terminal sensor nodes. To solve the problem, we also designed and implemented an unmanned ship equipped with terminal sensor nodes to facilitate the deployment of terminal nodes. Through assessment of the complexity of the network topology and energy consumption, it was demonstrated that LoRaWAN is superior to the traditional ZigBee network to deploy a monitoring network in the island aquaculture environment. Our tests also showed that this actual system can effectively achieve the long-distance data acquisition and transmission of the environment information for the island aquaculture with low power consumption. Network lifetime and transmission reliability tests showed that the network has an effective lifetime of 2.4 years when terminal sensor nodes use the 3.7V/4200 mAh battery and the upload period is 30 minutes. Within 800 m for the communication distance, the node packet loss rate is less than 3.6% when the transmission power is 20 mW. Furthermore, it is found that the received signal strength fluctuates significantly within 24 hours. We investigated that at night, the received signal strength decreases significantly and the quality of the communication link reduces. It showed that the island environment of the aquaculture area has a great impact on the communication quality of the LPWAN system. In particular, the temperature changes at night and the wind enhancement can lead to rapid decay of the LoRa signal and the quality of the network communication drops significantly. According to the test in the island environment, the system can effectively transmit data from the terminal sensor nodes and reflect the information of island aquaculture environment.