The end of the 20th century and the beginning of the 21st century in the Czech Republic were characterized by frequent extreme water cycle fluctuations, i.e. the occurrence of increased incidences of flood and drought events. Drought occurs irregularly in the Czech Republic during periods with low precipitation amounts. The most noteworthy droughts with significant impact, especially to agriculture, occurred in the years 2000, 2003, 2007, 2009, 2012, 2014 and 2015. A significant increase in frequency and length of drought periods was detected in future climate projections based on the latest model outputs, such as from the Euro-CORDEX 0.11° resolutions for the European area. For these model experiments, the following greenhouse gas emissions scenarios were used: Representative Concentration Pathway (RCP) 4.5 (milder scenario) and RCP8.5 (pessimistic scenario). Since the climate models suffer from potentially severe biases, it is necessary to statistically correct their outputs. For this purpose, a suitable reference dataset was prepared, based on quality-controlled, homogenized and gap-filled station time series. The correction method applied was based on variable correction using individual percentiles. From the corrected model outputs, selected extreme indexes with respect to drought analysis were calculated. From the results, it follows that we can expect both an increase in air temperature and in precipitation (with increased amounts per event), as well as an increase in other extremes with the capability of inducing drought (number of tropical days, heat waves, etc.).