Recently, new rapid methods of sensory analysis have been used as alternatives to common methodologies. The free Sorting test is a single session similarity-based method. Samples are presented simultaneously to the assessors and they are required to sort them into groups according to their similarity. Finally, the assessor must give each group its defining characteristics. The aim of this work was to study the application of free Sorting test methodology in sensory evaluation of a complex product: coffee. Six commercial coffee, common (3) and gourmet (3), were selected and prepared according to manufacturer's indication. The test was applied to 70 coffee regular consumers. Consumers received all the samples and were asked to group the samples according to their characteristics. A principal component analysis (PCA) was done using a correlation matrix with the means of the samples. Hierarchical cluster analysis (HCA) was also performed to find groups of samples with similar sensory characteristics, considering Euclidean distances (dissimilary) and Warld's techniques (agglomeration method) and automatic truncation. According to PCA for Sorting, two samples (Gourmet 2 and 3) were grouped for acidity, weak aroma, weak and weak color. These samples were discriminated from the others. The coffees (Gourmet 1, Common 4, 5 and 6) were responsible for pleasant aroma, strongness, appearance, color, flavor, bitterness. Bitterness and aroma were the terms most commonly cited with 141 and 80 citations, respectively. Other common characteristics were acidity, color, strong taste and weak. Results obtained with HCA showed four groups with distinct characteristics of coffee. This study concluded that the Sorting is not an ideal method for the common coffee consumer to characterize coffee, a food matrice with multiple and complex sensory characteristics.