Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a useful technique for the identification of bacteria on the basis of their characteristic protein mass spectrum fingerprint. Highly standardized instrumental analytical performance and bacterial culture conditions are required to achieve useful information. A chemometric approach based on multivariate analysis techniques was developed for the analysis of MALDI data of different bacteria to allow their identification from their fingerprint. Principal component analysis, linear discriminant analysis (LDA) and soft independent modelling of class analogy (SIMCA) were applied to the analysis of the MALDI MS mass spectra of two pathogenic bacteria, Escherichia coli O157:H7 and Yersinia enterocolitica, and the non-pathogenic E. coli MC1061. Spectra variability was assessed by growing bacteria in different media and analysing them at different culture growth times. After selection of the relevant variables, which allows the evaluation of an m/z value pattern with high discriminant power, the identification of bacteria by LDA and SIMCA was performed independently of the experimental conditions used. In order to better evaluate the analytical performance of the approach used, the ability to correctly classify different bacteria, six wild-type strains of E. coli O157:H7, was also studied and a combination of different chemometric techniques with a severe validation was developed. The analysis of spiked bovine meat samples and the agreement with an independent chemiluminescent enzyme immunoassay demonstrated the applicability of the method developed for the detection of bacteria in real samples. The easy automation of the MALDI method and the ability of multivariate techniques to reduce interlaboratory variability associated with bacterial growth time and conditions suggest the usefulness of the proposed MALDI MS approach for rapid routine food safety checks.