The development of a multivariate model for fiber content prediction in sugar cane has been investigated and can be offered as an alternative to the wet methods of analysis. The importance of fiber in bagasse of sugar cane is directly linked to the production of sugar and alcohol performances and to payment systems. The analytical method investigated here was the Laser-Induced Breakdown Spectroscopy (LIBS) combined with a Partial Least Squares (PLS) chemometric tool that can achieve the best model to predict fiber content in sugar cane. A total of 6287 spectra of raw samples of sugar cane bagasse was evaluated. The optimum conditions of operation of the equipment for the acquisition of the most sensitive LIBS signals were tested through Design of Experiments (DOE). (C) 2015 Elsevier Ltd. All rights reserved.