Electrical rotating machines are among the most common assets used in industry. In railways applications these devices are present in fixed and rolling stock systems, such as turnouts and traction components. Condition based maintenance (CBM) of rotating machines may significantly improve the availability of critical railway assets. Moreover, by efficiently assessing the state of health of targeted components, it becomes possible to introduce advanced asset management strategies for life cycle cost optimization. In comparison with traditional maintenance approaches, health monitoring enables better maintenance scheduling, fleet size optimization and maintenance costs reduction. CBM applied to rotating machines has been actively studied by many researchers in a wide variety of fields such as: signal processing, anomaly detection, failure diagnostic and failure prognostics. However, there is still a considerable gap between the methods studied in research and the ones successfully applied in industry, and especially in the railway field. This paper discusses the challenges and opportunities for application of CBM methods to electrical rotating machines in railway applications. For the purpose of illustration, a case study focusing on traction motor bearings is considered. Time domain and frequencydomain signal processing techniques are employed to extract features from bearing degradation data. The data analyzed in the present study have been obtained in a bearing test bench and during a test conducted on a real traction motor used in trains. The results of the considered methods are discussed and future research directions are suggested.