Groundwater is an important source of water supply for irrigation in Tunisia and many parts of the world. Therefore, this study aimed to delineate the groundwater levels in Mahdia that declined progressively due to intensive use in agriculture. In this study, statistical analysis of collected data, semivariogram model selection, ANOVA technique, removing trend component from observed data, and zoning maps using geostatistical tools in the Jeostat software were performed. Groundwater level data from the rainy period in 2008 and 2018 were used to compare the spatial changes in one decade and determine the overexploited areas. The experimental semivariograms enable us to show the trend component. Then, the ANOVA technique should be used to model the trend and remove it from the observed data. The spatiotemporal variation of groundwater table elevations are mapped by summing up residuals and quadratic trend model maps. These resulting maps show the groundwater table decline in many areas when rainfall is less than average for several years. The drought effect and groundwater overexploitation have damaging consequences and could reduce the groundwater level by 5 m in some areas.