The study of agricultural soil compaction is of great relevance, due to its negative effects on crop development and productivity. A quickly and easy way to identify soil compaction is throughout penetration resistance (PR). The aim of this study was to define an approach to adequately characterize the vertical and horizontal variability of agricultural soil compaction, using multivariate analysis techniques. Soil penetration resistance was measured before tillage and after crop harvesting a field grown with maize (Zea mays), using a 32 point regular grid (25 m X 25 m), up to 60 cm of depth. To identify compacted areas and layers, multivariate analysis techniques were used. The hierarchical cluster analysis, considering the RP measurements by sampling point, allowed drawing contour maps with defined compacted areas. Meanwhile, when such an analysis was performed by depth, together with the principal component analysis, three soil layers were identified being the middle layer the one with the highest values of PR in both sampling periods. The approach performed, allowed characterizing the vertical and horizontal variability of PR in the field under study.