<p>Direct measurement of groundwater recharge rates in the field is nearly impossible, so various hydrological methods have been used. Especially in areas where long-term groundwater level observations are conducted, the groundwater recharge rate can be estimated by applying the hybrid-WTF (Water table fluctuation) method. In South Korea, thousands of groundwater monitoring wells are operated for various purposes, enabling the estimation of groundwater recharge rates using the hybrid-WTF method. Therefore, this study aimed to predict the average groundwater recharge rate of sub-watersheds by utilizing the relationship between the recharge rates at observation points and the characteristics of each site. Groundwater recharge rate prediction was performed using the regression tree method (significance level 0.05). For a total of 3,886 observation wells, groundwater recharge rates (average 14.66%) calculated by the hybrid-WTF method using groundwater levels were set as the output variable. The topographic elevation, slope, stream, geology, soil, land cover, well density, and rainfall of each site were set as input variables. A regression tree model that classifies into a total of 30 groundwater recharge rates has been finally developed. The input variable with the most significant impact on classification was topsoil (predictive variable importance: 0.59), which correlates with the presence of a soil porosity factor in the hybrid-WTF equation. After applying a regression tree model to all 30,936 grids in Chungcheongnamdo province to estimate the groundwater recharge rate, the average recharge rate was found to be about 12.8% of the precipitation. The province is composed of a total of 609 sub-watersheds, with the smallest sub-watershed being about 5 km² and consisting of 20 grids. The REC (Representative elementary count) analysis indicated that the average groundwater recharge rate predicted in more than 25 random grids is almost identical to the original average recharge rate. Therefore, considering the number of grids per sub-watershed, the recharge rate for each sub-watershed can be calculated as the average of the grids’ recharge rates. Ultimately, the groundwater recharge rates for 609 sub-watersheds in the Chungcheongnamdo province were predicted to be approximately 27.53% at maximum and about 2.84% at minimum. While it is very challenging to accurately determine the groundwater recharge rate for specific sites due to their diverse characteristics, estimating the groundwater recharge rate over a certain scale is feasible using big data models, and it is hoped that this will lead to more reliable policies.</p>