The meteorological information decision support system plays an important role in the process of meteorological disaster prevention and reduction, but the monolithic architecture of a meteorological information decision support system is difficult to meet the needs of scientific decisionmaking and intelligent disaster reduction, and there are deficiencies in scalability and reusability. The microservice architecture based on Spring Cloud has the characteristics of low coupling and easy expansion, and has been widely used in large and mediumsized enterprises. In this paper, through the research on the key technologies of microservice architecture, three reconstruction strategies of meteorological information decision support system architecture are proposed. In view of the limitations of meteorological information decision support systems, combined with the characteristics of microservice architecture, the framework of meteorological information decision support systems based on microservice is designed, which is divided into the core layer, business logic layer and presentation layer. The business logic layer is mainly reconstructed and realized, and the reconstructed business functions are applied to the meteorological disaster prevention and mitigation monitoring management platform, which improves the scalability and reusability.