摘要： The ever-increasing speed of exchange of ideas, information, and culture allows contemporary art to be in constant growth, especially concerning the choice of artistic materials. Their characterization is not only crucial for the study of artistic techniques but also for research into the stability of the material and, consequently, the best preservation practices. For this aim, an analytical method should have the advantages of not requiring sample preparation, performing superficial micro-analysis, and obtaining detailed spectral information. For this study, laser-induced breakdown spectroscopy (LIBS) was employed. It was used for the identification of modern paints composed of inorganic pigments and organic binders, such as acrylics, alkyds, and styrene-acrylics. Principal component analysis (PCA) was used to classify the different pure materials, above all, the polymeric binders. To distinguish the paint mixtures, whose LIBS spectral results were more complex due to the pigment/binder interaction, a statistical method recently employed in the cultural heritage field was chosen, namely, random decision forest (RDF). This methodology allows a reduction of the variance of the data, testing of different training data sets by cross-validation, an increase of the predictive power. Furthermore, for the first time, the distribution of different inorganic pigments and organic binder materials in an unknown sample was mapped and correctly classified using the developed RDF. This study represents the first approach for the classification of modern and contemporary materials using LIBS combined with two different multivariate analyses. Subsequent optimization of measurement parameters and data processing will be considered in order to extend its employment to other artistic materials and conservation treatments.