Some key technologies and methods for the pre-processing of global atmospheric reanalysis data based on global real-time meteorological data are developed for China's first generation global atmosphere and land surface reanalysis system (CRA) and described in this paper. The key technologies mainly include: the data sorting technology that can realize rapid analysis and classification of a large number of real-time observations, the information extraction technology that can realize parallel decoding of multi-types of meteorological observation elements, and the custom template unified coding technology that can meet the requirements of assimilated data formats and the effective classified data storage. These key technologies have been applied in the National Meteorological Information Center in June 2019, providing stable real-time global observation data support for CRA.