Mid-infrared fiberoptics reflectance spectroscopy (mid-IR FORS) is a very interesting technique for artwork characterization purposes. However, the fact that the spectra obtained are a mixture of surface (specular) and volume (diffuse) reflection is a significant drawback. The physical and chemical features of the artwork surface may produce distortions in the spectra that hinder comparison with reference databases acquired in transmission mode. Several studies attempted to understand the influence of the different variables and propose procedures to improve the interpretation of the spectra. This article is focused on the application of mid-IR FORS and multivariate calibration to the analysis of easel paintings. The objectives are the evaluation of the influence of the surface roughness on the spectra, the influence of the matrix composition for the classification of unknown spectra, and the capability of obtaining pigment composition mappings. A first evaluation of a fast procedure for spectra management and pigment discrimination is discussed. The results demonstrate the capability of multivariate methods, principal component analysis (PCA), and partial least squares discrimination analysis (PLS-DA), to model the distortions of the reflectance spectra and to delimitate and discriminate areas of uniform composition. The roughness of the painting surface is found to be an important factor affecting the shape and relative intensity of the spectra. A mapping of the major pigments of a painting is possible using mid-IR FORS and PLS-DA when the calibration set is a palette that includes the potential pigments present in the artwork mixed with the appropriate binder and that shows the different paint textures.