Deoxynivalenol and fumonisins pose a health concern and have economic consequences, so the European regulation CE 1126/2007 dictates the maximal content allowed in cereals. The direct measurement of mycotoxin content using the established method is not only time-consuming and tedious, but also destructive and cannot be used in a silo. Alternative tools such as infrared spectroscopy are therefore being studied. For the present investigation, spectral data collected from maize kernels contaminated naturally by mycotoxins were studied to predict the risk of deoxynivalenol and fumonisins. Discriminant models were used to create and identify batches that satisfy regulations for animal or human consumption.