In order to deal with the ill-posed problem of material parameter identification for rockfill materials, a procedure based on machine learning is proposed. In this investigation the results of a comprehensive set of tests on rockfill materials are examined. The instances, categorized in accordance with their particles shape (angular / rounded), gradation characteristics and relative density, are analyzed using Regression Trees, a machine learning tool that deals with the construction and study of algorithms for learning from data. The emphasis of this study is on determining resistance and deformability behavior of the rockfill. It is shown that high confining stresses and particle breakage phenomenon are found as the driving factors of the behavior of the materials.