Abstract: The paper aims to analyze rainfalls of typical flood years over the past 30 years in representative counties of southern, central and northern Guangxi. The results showed that: the linear correlation between the effective rainfall of the waterlogging disaster day and that of the 6 days before flood was most obvious, reached 95% confidence level. Based on this finding as well as previous studies and disasters in history, the grade index of farmland waterlogging was constructed by the sum of the effective rainfall of the waterlogging disaster day and that of the 6 days before the flood. Taking the remote sensing information of the spatial distribution of double-cropping rice in Guangxi as the waterlogging disaster bearing body, smart grid point rainfall forecast products were embedded in and mixed with the effective rainfall raster maps of real- time rainfall from township automatic observation stations, to realize the quantitative assessment of double-cropping rice before waterlogging disaster in Guangxi. This research can provide certain references for government institutions to carry out disaster prevention and mitigation, and promote direct and smart meteorological service for agricultural production.