Abstract: To address issues such as lack of specificity and targeting in traditional meteorological warnings, and generalized services scope, a loquat meteorological disaster warning technology based on WeChat service accounts has been proposed and implemented．This technology focused on key phenological stages during loquat growth cycle．By integrating real-time meteorological observation data, grid forecast data, and other information, a multi-factor meteorological disaster indicator discrimination model was constructed．A multi-factor weighted evaluation algorithm was employed to achieve automatic disaster risk identification and tiered warnings．After being manually reviewed and verified, warning information was delivered to subscribing farmers through WeChat service account interface．Since its implementation in 2020, this technology has enabled precise and targeted delivery for loquat meteorological disaster warnings, with warning results largely consistent with actual disaster occurrences．Compared to traditional regional broadcast warnings, it has significantly improved service specificity and user-friendliness, enhancing refined level and practical usability of meteorological services．