Abstract: Extreme meteorological disasters such as droughts, floods, heat stress, and low-temperature frost damage are increasing in frequency, spatial extent and severity in the context of global climate change. Additionally, shifts in modern agricultural production systems and the emergence of new technologies such as artificial intelligence and big data present novel opportunities for the development of agricultural meteorological services. Convenient and accurate agricultural meteorological services can provide critical support for safeguarding food security and enhancing disaster prevention and mitigation efforts. To further enhance the application capability of national agricultural meteorological services, the new version of China Agricultural Meteorological Service System (CAgMSS3.0) is under development based on the existing CAgMSS2.0 framework and is integrated with Meteorological Big Data Cloud Platform (Tianqing) of China Meteorological Administration. CAgMSS3.0 utilizes Tianqing cloud servers for the deployment of its basic data and algorithms. Compared with CAgMSS2.0, several new modules are introduced, such as crop meteorological suitability index, annual agroclimatic evaluation and prediction, all-weather crop growth condition monitoring and analysis via optical and microwave remote sensing, agricultural meteorological disaster index, and grid-based agricultural meteorological disaster monitoring and prediction. Furthermore, CAgMSS3.0 has improved soil moisture monitoring and evaluation by integrating machine learning with multi-source data fusion. It also incorporates advanced meteorological forecasting technology for the occurrence and development of agricultural pest and disease, an interactive national-provincial agricultural weather prediction framework, and refined methods for agricultural climate zoning as well as agricultural meteorological disaster risk zoning. This system significantly enhances the operational capacity of national agricultural meteorological services through its application. Nevertheless, CAgMSS3.0 has some limitations. First, functional modules currently lack integration of global agricultural meteorological monitoring and forecasting components. Second, emerging domains such as climate quality monitoring and forecasting for agricultural products, as well as agricultural meteorological financial and insurance services require further development in the system. Third, the application of cutting-edge technology, especially AI-driven decision support in agricultural meteorology, remains undeveloped. Future iterations of agricultural meteorological service system are expected to be incorporated into a new-generation weather business integration platform structured around an "intelligent core". Meanwhile, a large-scale model based on "AI + mechanism model" will be developed for crop growth simulation and intelligent agricultural meteorological services. These improvements are anticipated to facilitate more efficient, accurate, and intelligent agricultural meteorological services.