It is very important to study the extreme weather variations, to define the threshold of extreme rainfall event and to evaluate the risk of internal water-logging for highly urbanizing areas in Southern China for improving the forecasting and early warning schemes in the command system, and dealing with the extreme rainstorms and water-logging disasters in the courses of climate change and rapid urbanization. Taking Shenzhen as an example, based on data collection of the rainstorm events in Shenzhen from 2000 to 2015 and the corresponding rainfall data from several sites, the fuzzy identification method and PCI were used to analyze the rainfall pattern and rainfall concentration in the region. Also, the Shenzhen extreme rainfall threshold has been analyzed and calculated using the percentile method. The results showed that the single-peak type rainfalls were main rainfall type which caused water-loggings of Shenzhen; the main cause of waterlogging was uniform heavy rainfall; the highest rainfall threshold was located in the Dapeng New Area and the spatial distribution of rainfall thresholds decreased from the east to the west; the extreme rainfall threshold of Shenzhen was the rainfall with frequency of 1%.