Abstract: The improvement of the safety and the availability of the complex industrial systems passes by the detection and the localization of the failures of their components. A failure can be a total breakdown of an element of the system or a modification of its characteristics sufficiently significant to deteriorate the total operation of the installation. Safety is affected when a failure occurs and that it is not detected, in addition a not known stop due to an alarm (false alarm) decreases the availability of the systems. The purpose of the monitoring systems are to detect and identify the failures which can affect the performances and the safety of an installation with a minimum of false alarm and within times weakest possible and this starting from information available on a system (order, measurement and a model of behavior). This study consists with the application of the method of estimate parametric (isolation resistance of the armature and the inductor) for the monitoring of the physical parameters of an electric drive. Various approaches of defects monitoring are mentionned. This research treats also the motors with D.C. current as well as the various possible defects and their remedies. Our essential contribution relates to the proposal of algorithm detection based on the method of least squares, which is validated by simulation. In addition, it would be interesting to study, from the new point of view, other parametric methods of estimate of which that of recursive least squares for the application on line in order to detect the anomalies in all the types of motors.