This paper proposes the design idea and overall framework of the accurate and intelligent push platform for early warning information based on the Jiangxi Early Warning Release WeChat service, and builds a operational platform consisting of the operational support layer, data layer, service layer, and operational application layer to achieve the realtime release of the gridbased proximity prediction information and the targeted release of earlywarning information based on location. The earlywarning service model has shifted from “oneway push” to “twoway interaction.” The study applies a number of key technologies to support the intelligentization of the platform, including the core functions and technologies of the platform, such as LBS (LocationBased Service)based realtime warning, approaching rain conditions, early warning customization, targeted group sending, and background management. The platform is easy to operate and to use and provides users with tailored earlywarning services in real time.