The present paper aims to provide a metabonomic model for discrimination of soybeans according with their biochemical status as function of exposition, or not, to gamma irradiation. Herein, we used soybeans from five different cultivars supplied by Embrapa Soja (Londrina-Paraná-Brazil). The samples were exposed to doses of 1 and 5 kGy. We obtained 1H nuclear magnetic resonance (NMR) spectra from the chloroform extract of grains and then performed the chemometric processing using exploratory principal compo- nents analysis (PCA) and partial least square e discriminant anal- ysis (PLS-DA) formalism. The model was constructed and further validated to provide a powerful method for the classification of irradiated soybean seeds.