Multiscale analysis of precipitation series in the Yunnan-Guizhou Plateau is essential for the comprehensive management of karst region watersheds, rational utilization of water resources, and regional sustainable development. The complete ensemble empirical mode decomposition(CEEMD) algorithm combined with fast Fourier transform (FFT) was utilized for feature extraction. Precipitation observation data, collected from 60 stations over 20 years with a 12-hour interval, were used to identify periodicity in the precipitation series of the Yunnan-Guizhou Plateau. The results show that the CEEMD algorithm effectively extracts characteristic components, such as monthly, semi-annual, and annual cycles from the precipitation series. Analyzing the positions of various meteorological stations reveals that stations in close geographical proximity exhibit similar periodic variation trends. When compared with the Köppen-Geiger climate classification map of the Yunnan-Guizhou Plateau, the results align with regional climatic characteristics, validating the applicability of this decomposition method in climate statistics. How actual precipitation series respond to various climatic driving factors under different periodic characteristics was further explored. Using the autocorrelation function(ACF), the lag and correlation was examined between precipitation series and climatic oscillations such as the quasi biennial oscillation(QBO), the Niño3.4 index, and tropical cyclones. It is found that the regional precipitation series are influenced by both El Niño-Southern Oscillation(ENSO) and QBO. Specifically, 48.3% of the station intrinsic mode functions (IMFs) exhibit a bimodal annual cycle. By comparing extreme precipitation events with tropical cyclone precipitation events, it is discovered that the bimodal annual cycle in the Yunnan-Guizhou Plateau's precipitation series is influenced by tropical cyclones.