This paper presents a predictive maintenance technique being used at CST, the diagnosis of AC rotating electrical machines. These diagnosis techniques are based on variables monitoring that indicate the insulation contamination and degradation levels, making possible to assess the machine remaining lifetime, optimizing the maintenance planning and reducing the impacts over production and company assets. A case study of overhaul maintenance, based on the result of a diagnosis recommending the motor stator re-wedging, confirm the diagnosis techniques efficiency.