The present work aims to develop an automatic system to identify the maturation degree of iron ore pellets, involving techniques of microscopy and digital image processing. Iron ore pellets, as well as sinter and lump ore, are the basic iron-bearing burden used in ironmaking. In one of the steps of their production, the pellets acquire different microstructural characteristics, the socalled Maturation Degree, which influences directly in the reduction processes. Thus, microstructural characterization is an important step in quality control of the material. The microstructure evolution of pellets for blast furnaces was divided into four classes: A, B, C and D. Attributes that describe intrinsic characteristics of each class were extracted from optical microscope images and were used to train an automatic classifier. Three dimensionality reduction techniques were tested to optimize the classification system: exhaustive search, Principal Component Analysis and Fisher’s Linear Discriminant Analysis. Three types of classifiers were tested (Quadrático, Mahalanobis and Linear). The highest global success rate was above 90 per cent. This indicates that this methodology is effective for automatic classification of the Maturation Degree in blast furnace pellets.