Drought is one of the most severe natural hazards with a significant impact on water resources. Droughts affect the sustainability of water resources and may result in environmental degradation of a region. The appropriate drought index is playing an important role on drought assessment. This paper addresses the best-fitting distribution of monthly water balance (WB), and the spatiotemporal characteristics of drought in Guizhou Province, China, during 1960–2017. The best-fitting distribution of monthly WB calculated using three potential evapotranspiration (PET) estimation methods was determined based on the goodness-of-fit test. The Standardized Precipitation Evapotranspiration Index (SPEI) at various timescales was computed based on the optimal distribution. The modified Mann-Kendall method, run theory, and inverse distance weight (IDW) interpolation method were used to investigate spatiotemporal evolution of drought. Over result revealed that the common log-logistic distribution was not the most suitable distribution for WB in Guizhou Province and the gen. extreme value distribution is optimum fit for the WB series calculated based on different methods for 76.47% of the stations. Moreover, different calculation methods seem to have little influence on the fitting results. The drought shows an upward trend in spring and autumn and a downward trend in summer and winter during 1960–2017; the increasing magnitude of summer and autumn drought was greater in the northeastern and northwestern Guizhou Province. The annual droughts are significantly increasing (p = 0.05) at Anshun and Panxian. The drought frequency showed a decreasing trend from light drought to severe drought in Guizhou Province, and the drought intensity was relatively more severe during 2010–2017 than other periods. These findings are helpful for water resource management in Guizhou Province and indicate that we should pay more attention to the drought mitigation of the higher-elevation regions.