The objective of this research was to develop a mathematical and statistical model for long-term prediction. The Extreme
Value Theory (EVT) was applied to analyze the appropriate distribution model by using the peak-over-threshold approach
with Generalized Pareto Distribution (GPD) to predict daily extreme precipitation and extreme temperatures in eight
provinces located in the upper northeastern region of Thailand. Generally, each province has only 1–2 meteorological
stations, so spatial analysis cannot be performed comprehensively. Therefore, the reanalysis data were obtained from the
NOAA Physical Sciences Laboratory. The precipitation data were used for spatial analysis at the level of 25 square
kilometers, which comprises 71 grid points, whereas the temperature data were used for spatial analysis at the level of 50
square kilometers, which includes 19 grid points. According to the analysis results, GPD was appropriate for the goodness
of fit test with Kolmogorov-Smirnov Statistics (KS Test) according to the estimation for the return level in the annual
return periods of 2 years, 5 years, 10 years, 25 years, 50 years, and 100 years, indicating the areas with daily extreme
precipitation and extreme temperatures. The analysis results would be useful for supplementing decision-making in
planning to cope with risk areas as well as in effective planning for resources and prevention.