The traditional pest monitoring methods is a labor-intensive and time-consuming work which can't meet the requirements of modern precision agriculture. In order to realize the dynamic monitoring and forecasting of vegetable pests and provide information support for the development of precision agriculture, it has become an inevitable trend to realize the fast, safe and reliable transmission of pest information collection. In this paper, a comprehensive monitoring and early warning system of vegetable pest based on internet of things (lOT) is designed. The system includes farmland environment data acquisition system (EDAS), vegetable pest image data acquisition system (VPIDAS) and vegetable pest monitoring and early warning platform. The EDAS uses commercial environmental sensors to collect farmland environmental data. The VPIDAS is self-developed, this system combines the technologies of machine vision, telematics and solar power technology, and can obtain the image data of vegetable pests for a long time. The Vegetable pest monitoring and forecasting platform is used to manage the data acquisition system, and analyze the collected data to provide information support for forecasting and early warning of vegetable pests. Users can interact with the platform via Internet with a PC, management of data collection, analysis of the collected data and prediction of vegetable pests early warning. A series of real test-beds show that the monitoring system can work steadily and reliably, and has wide application prospect.