Raman micro-spectroscopy was applied to compile a large-scale database of Raman spectra of single Bacillus endospores and to calculate classification functions, which were trained to discriminate between endospores of 66 strains from 13 Bacillus and Bacillus-related species including B. anthracis. The developed two-stage classification system comprising two support vector machines and one linear discriminant analysis classifier was then challenged by a test set of 27 samples to simulate the case of a real-world-scenario, when "unknown samples" are to be identified. In the end, all 27 test set samples including six B. anthracis strains were identified correctly. The samples thereby covered a diverse selection of species within the phylogenetically broad Bacillus genus and also included strains, which were not incorporated in the database before. All of them were correctly identified on the species level with accuracies between 88 and 100%. The sample analysis itself requires no biomass enrichment step prior to the analysis and qualifies the presented Raman spectroscopic approach to be a rapid analysis system in term of Bacillus endospore typing.