In the era of Industrial Internet of Things (IIoT) and Industry 4.0, complex electromechanical systems will be equipped with a variety of sensors providing new opportunities for the development of Health Monitoring and Management Systems focusing to the optimum exploitation of available information in order to maximize the performance of machinery. One of the 10 EU Commission priorities since 2015 is the “Digital Single Market” (in order to ensure that Europe’s economy, industry and employment take full advantage of what digitalization offers) while world technology leaders such as Rolls Royce, Siemens and GE work towards the development of special digital platforms (R2Data®, MindSphere® & Predix®). Focusing towards the increase of production reliability and safety as well as on the reduction of cost, there is an ever increasing industrial need not only for accurate, early, on time and online fault detection and diagnosis with minimum/optimum number of false alarms and missed detections but also for a robust, early, accurate and on time estimation of the Remaining Useful Life (RUL) of the defected components, within a confidence interval, independent of the operating conditions. Prognostics and Health Management (PHM) is an emerging engineering discipline, linking the failure mechanisms to the system life cycle management, but is still an Achilles’ heel in Condition Based Monitoring being still immature. Prognostics is extremely important for safety critical components and therefore the first prognostics applications are focused on aerospace vehicle applications (NASA, GE, Rolls Royce, Pratt&Whitney), on electronics & battery applications (related mainly to aerospace) and recently on industrial applications such as paper making machines. Moreover the existing Prognostic and Health Monitoring (PHM) techniques have not yet achieved the level of accuracy needed for systems operating under varying operating conditions and have not yet found massively their position in the industrial world. A key technological barrier is the absence of real (mainly vibration) measurement data from the industrial field in sufficient quantities (enough to assure the correct training and validation of Machine Learning algorithms), captured under different operating conditions and presenting different failure modes and abnormalities, as industries seldom allow their machines to run to failure. Additionally the physical degradation tests, even in their accelerated version, present long duration and extremely high cost without reassuring the natural development of different failure modes. Therefore, to overcome the abovementioned limitations, the core target of the project is the development of a Digital Twin approach for Health Monitoring and Predictive Maintenance of transmissions, consisting of bearings and gears. The major objectives of the project are: a) the development of a Digital Twin of physical component/system in operation, to support data driven prognostic methodologies by simulations and artificial data generation, trying to solve the classical problem of unavailable, sparse or truncated data and b) the proposal of advanced prognostic techniques for the estimation of Remaining Useful Life of mechanical components/systems based on the combination of the Digital Twin perspective with Machine Learning and Parameter Estimation techniques.