The global urbanization process is advancing rapidly, and urban flood disasters occur frequently, which has become a common urban problem faced by all countries in the world, especially China and other developing countries. Analyzing the time evolution law and development trend of flood phenomena has important reference value for taking scientific policy measures to effectively prevent and control flood disasters. Based on the daily meteorological data of Guiyang from 1961 to 2020, using SPEI index, Mann Kendall test method and wavelet transform, this paper analyzes the flood and waterlogging change rule under multi-scale time series in recent 60 years. The results show that: ① The monthly and seasonal precipitation intensity in Guiyang is high, the frequency of rainstorm is high, the monthly precipitation intensity is concentrated in the heavy rain and above, and the seasonal precipitation intensity is concentrated in the heavy rain and above, so extreme precipitation events are easy to occur, causing urban flood disasters; ② At multiple time scales, SPEI values have different sensitivity to precipitation, temperature and other meteorological factors. The sensitivity of daily scale is the strongest, while that of other scales tends to be flat. At the same time, the proportion of flood events corresponding to large scales gradually decreases. ③ The year of flood in Guiyang City accounts for 21.67% of the total year. SPEI-12 shows an upward trend. The mutation of SPEI-3 occurred in the 1960s. After the mutation, it shows an upward trend and a significant upward trend in autumn and winter, indicating that seasonal flood has an enhanced trend; ④ The cycle change characteristics of SPEI-1 are as follows: the cycle is 23 months under the condition of 33 month main cycle; The period under the condition of 14 month main period is 9 months.