This dataset compiles the georeferenced soil sampling tables produced within the Master’s research in Agricultural Engineering (UNICAMP). It contains the soil fertility data collected in three agricultural areas located in São Paulo State, Brazil. The dataset includes sampling tables corresponding to three sampling intensities (1, 0.5, and 0.2 samples per hectare), representing different density scenarios evaluated in the study. Each table contains geographic coordinates (UTM) and soil fertility attributes (P, K, CTC, pH, V%, and clay content). These data were used as input for the generation of soil fertility maps using six spatial interpolation methods (IDW, TPS, OK, RK, RF, and SVM) and for the development of variable-rate fertilizer prescription maps based on regional agronomic recommendations. The interpolation procedures and prescription criteria are fully described in the associated Master’s thesis and related publications.