This thesis presents the development of a new technique for fault diagnosis in complex systems called the Latent Nesting Method. This technique, created in response to some predecessors dissertation which provide the diagnosis in relation to the problems and concepts of combinatorial explosion, modular decomposition and intermittent faults with this giving way to new diagnostic and monitoring techniques geared toward predictive maintenance systems such as: monitoring the state or condition monitoring. In relation to this new technique of nesting latent modeling tools are used such as Petri Nets and some of its most important extensions such as the Coloured and hybrid type which give the method its theoretical basis and mathematical. In the context of diagnosis, the methodology offers a new possibility to detect different types of failures in complex systems either because of their complexity in terms of signals and variables, or its operation and application, regardless of their nature is discrete, continuous or hybrid as in most complex cases. The different characteristics of the methodology, as its design and implementation allow us to take a step forward in solving diagnostic problems in a simple and fast computational models solving the classic paradigm of diagnosis: detection, isolation and identification. Along with the theoretical formulations and models made in the thesis, we present the experimental results and simulations from two totally different complex systems in their application, which see the feasibility of the method and its practical implication. These two applications are geared toward diagnosing a failure in wind turbines (complex systems of many different signals), and the second to the diagnosis of failures in IGBT power semiconductors (complex systems in their operation). Discussions and demonstrations presented in this thesis include a detailed description of the implications that led to the achievement of this new method, a theoretical basis for its development and applications of experimental and test operation.