Soil erodibility is one of the most important factors in understanding the erosive process. In view of the need to explore methods for determining the values of erodibility by simulated rainfall, the objective was to evaluate, through the tools of multivariate statistics, the erodibility of Latosols from Alagoas influenced by the physical, chemical and stability attributes of aggregates. The research activities were carried out at the Arapiraca Campus, of the Federal University of Alagoas (Universidade Federal de Alagoas), where Latosols were used in the suborders Red, Yellow and Red-Yellow, collected in two of the three geomorphological regions of the state of Alagoas. The prediction of correlation metrics shows that the cluster grouping indicates that the best correlations found were between the variables: clay and clay dispersed in water (r = 0.94), lime and sand (r = 90), and between the Water Erosion Prediction Project and the erodibility factor in midgrooves (r = 1.00). These strong positive correlations are proven through scatterplots, confidence regions, dendrogram, density estimation, and ellipses. The multivariate statistics allow a better understanding of the behavior of the correlations between the physical, chemical and stability attributes of aggregates with the erodibility of Latosols of Alagoas, as well as demonstrating the relationship of the variables studied with any of the three suborders of soils of Alagoas (Yellow, Red and Red-Yellow Latosol).