In this paper, a pattern recognition (PR) method is used to provide the tracking and the diagnosis of a system. First of all, from measurements carried out on the system, features are extracted from current and voltage measurements without any other sensors. These features are used to build up a pattern vector, which is considered as the system signature. Then, a feature selection method is applied in order to select the most relevant features, which define the representation space. The decision phase is based on the "k-nearest neighbors" (knn) rule, associated with an evolution tracking of system using trajectory allowing a diagnosis not only of states defined in the training set, but also of the intermediate states. The appearance of a new operating mode is taken into account in order to enrich the initial knowledge base and thus to improve the diagnosis. This approach is illustrated on asynchronous motor of 5.5 kW with squirrel cage, in order to detect broken bars under any load level. The experimental results prove the efficiency of PR methods in condition monitoring of electrical machines