Profitable production demands continuously improved maintenance decision accuracy to reduce unplanned stoppages, prolong components/machines  life length and maintain production continuity. When CM, e.g. vibration level or oil temperature, value exceeds a warning level, it demands a clear understanding of what happened and how it will develop to avoid failures. It demands also reliable information, e.g. probability of failure, residual  life and when is the most profitable time for maintenance. Companies strive to reduce production cost through reducing non-value adding times, e.g. failures, for offering customers lower prices and generating additional competitive advantages. In this paper, a new tool (Prediction of Vibration and oil temperature, PreVib) for predicting the condition of a turbine, generator and gearbox is introduced,  discussed and tested. The solution offers a unique opportunity to achieve just in time dynamic and cost-effective maintenance by selecting the most profitable time for maintenance. PreVib is tested in a power plant. The major results: Failures can appreciably be reduced thanks the high precision of prediction ( + 90%), i.e. a negligible difference from the real measurement: PreVib does not demand extensive or additional data, and not replace available maintenance systems. The major conclusion is that applying PreVib, it is possible to reduce failures appreciably, prolong life length of components/equipment and perform profitable maintenance