echnological developments had represented a challenge for the industrial in order to improve the machinery and the productive systems, motivating an increase in their demand from community searching a stable economy and a reliable plant. Besides to decrease the environment and human risks. Therefore, the importance of maintenance area have increased because it allows to hold the systems health and its availability. The continued progress in sensing capabilities together the necessity of monitoring processes, have changed the industry paradigm, where the companies are migrating from traditional preventive maintenance strategies to conservative maintenance tasks, and they are incorporating predictive maintenance concepts, which are carried out only when these are required. In this context, speaking about machine fault diagnostics turns relevant and it can be defined as procedure of mapping the information obtained from a measurement space to a feature space. This mapping process is also called pattern recognition. Traditionally, pattern recognition is done manually with auxiliary graphical tools such as power spectrum graph, phase spectrum graph, cepstrum graph, AR spectrum graph, spectrogram, wavelet scalogram, wavelet phase graph, etc. However, manual pattern recognition requires expertise in the specific area of the diagnostic application. Thus, highly trained and skilled personnel is needed. Therefore, automatic pattern recognition is highly desirable. This can be achieved by classification of signals based on the information or features extracted from mechanical vibration signals. It is worth noting that vibration analysis is attractive in industry due to its low cost and the acceptable precision that it can reach using this technical tool, in comparison with techniques as acoustic emission and thermography. In consequence, this proposal covers only the issue related to digital processing of vibration signals. In Colombia, the industrial sector had seen the necessity of implement predictive maintenance programs, aiming to optimize the performance and the useful-life time of your machines. Hence, the vibration analysis as tool for machine diagnostics has generated all kind of expectations, because it has a high profitability with respect to cost/benefit. Therefore, the development of a methodology in machine diagnostics allows to increase the predictive and preventive maintenance in Colombian industry, entailing a major competitively in their products and processes. Regarding, the Signal Processing and Recognition Group focuses your activity on the research and development of stochastic haracterization, training and recognition systems, which are applied to machine diagnostics using the vibration analysis. Additionally, this doctoral proposal is framed into the research project titled “Desarrollo de un sistema piloto de mantenimiento predictivo en la lınea de propulsion de las lanchas patrulleras de la Armada Nacional empleando analisis de vibraciones e imagenes termograficas”, which was endorsed by Colciencias.