近年来，中国农业病虫草害问题日趋严重，农业病虫草害预测预报体系不完善，基本原因在于农业大数据的汇集能力、挖掘能力、决策能力不足，目前市场上已经有许多涉及病虫草害的信息平台，但都面临着种类划分不够统一、资源信息不够准确等问题。安徽省是农业大省，病虫草害问题尤为严重，为了推动安徽省病虫草害防控体系发展，本文旨在以病虫草害数字图像库为基础，构建大数据管理、分析、挖掘及可视化展示平台，实现数据资源的分布存储与处理，面向农业生产、管理决策和科技创新中的现实需求，开展系统框架、数据清洗、数据挖掘、知识发现、认知计算、数据建模等技术研究与产品研发，建设集管理、共享、创新应用及产品服务一体化的大数据平台。为从业者提供农业病虫草害识别和辅助诊疗、病虫草害预测预报、植保知识查询等植保信息精准化服务，突破时间、地域限制，利用互联网帮助从业者实时解决生产中遇到的病虫草害防治难题，降低经济成本，减轻作业强度，提高防治的时效性。最后，针对当前安徽省植保大数据平台的不足提出建议，未来要进一步补充平台所缺乏的遥感、气象、土壤等方面的信息化数据，并增加录入病虫草害的种类和数量，提高数据共享水平，优化数据分析技术，加强数据应用推广，完善数据安全保障，真正成为智慧农业的重要组成部分。

In recent years, China’s agricultural plant protection has become increasingly important, and the forecasting system of agricultural plant protection is not perfect. The basic reason is that the collection, mining, and decision-making ability of agricultural big data are insufficient. At present, there are many information platforms related to agricultural plant protection on the market, but they all face problems, such as insufficient classification of resources and inaccurate resource information. As a major agricultural province, the problem of plant protection in Anhui Province is particularly serious. To promote the development of a plant protection system in Anhui Province, this article uses the digital image library of agricultural plant protection as the bottom layer, and aims to realize the distributed storage and processing of data resources by constructing a platform for big data management, analysis, mining, and visual display. To meet the actual requirements of agricultural production, management decision making, and technological innovation, we conduct technical research and product research on a system framework, data cleaning, data mining, knowledge discovery, cognitive computing, and data modeling. We construct a big data platform integrating management, sharing, innovative applications, and services. We provide practitioners with accurate plant protection information, such as identification and auxiliary diagnosis of agricultural disasters, prediction of agricultural disasters, and plant protection knowledge. We overcome temporal and geographical restrictions using the Internet to help practitioners solve the difficult problems of plant protection in production in real time. This measure can reduce economic costs, reduce operation intensity, and improve the timeliness of prevention and control. Finally, we provide suggestions to address the shortcomings of current big data platforms for plant protection. In the future, it will be necessary to supplement the platform with further informational data that it lacks, such as data on remote sensing, meteorology, and soil; increase the number of diseases, pests, and weeds in the database; improve the data sharing level; optimize the data analysis technology; strengthen data application and promotion; and improve data security guarantees, become an important part of smart agriculture