Abstract: Currently, the national meteorological data service systems have basically met the demand of operational systems and scientific research. But, as the deep development of the meteorological operations, it's urgent to develop special data services to meet the specific requirements. For instance, the special data service system for climate monitoring is designed and implemented.The system is made up of one supporting database and four modules: Data-update monitoring, special dataset producing, data serving and task scheduling modules. The supporting database is in charge of fetching meteorological data from basic databases and storing information for the system operation. The data-update monitoring module detects the status of data updating with the strategies designed for each dataset, and triggers the special dataset producing module. The special dataset producing module produces datasets with scientific and practical methods, including analyzing, filtering, arranging, making statistics and standardizing of original data. The data serving module provides two ways for service: Data pushing and downloading via FTP. The task scheduling module dispatches the other modules according to business rules, realizes a complete operational flow perfectly by implementing three scheduling modes: Real-time scheduling, exception handling, and manual calling.The system is developed based on SSH architecture, DWR technology and configuration management. SSH involves three open-source products: Struts, Spring and Hibernate, which separately implement three layers of the system: Web representation layer, business logic layer and data access layer. The web representation layer provides a web platform for data manager submitting manual calling of producing datasets with the method of synchronous and asynchronous interaction, realized by using Struts and DWR. The business logical layer is based on Spring product, which implements all of the four operational modules, and implements the calling interface for the web representation layer. The data access layer provides data fetching interface by creating views of data in basic databases and encapsulating data access objects with Hibernate. In addition, the skills of configuration management, including the business configuring and the third-party software configuring, enhance the flexibility and scalability of the system.The system is a successful attempt on special meteorological data service. After being put into operation, it produces thirteen real-time updating special datasets, and serves climate monitoring systems very well. The specific dataset producing method, complete strategy of task scheduling, management skills of configuration, and the technical framework of the system provide a widely reference and promotional value for other special meteorological data service systems' design and implementation.