Abstract: Shouxian National Climate Observatory has accumulated a large amount of observation data so far. Taking the development process of 38 data sets involving surface, radiation, agriculture, atmospheric composition, radar and scientific experimental meteorology completed in 2019 as an example, this paper introduces the five specific steps of the research and development of long series meteorological scientific experimental data sets, namely, the determination of source data, quality control, entity data set production and processing, inspection and evaluation of data sets and compilation of data set description documents. The application of the datasets is discussed. The purpose is to further improve the development capability and standardization level of meteorological data products, and help the meteorological data sharing service.