Preventive maintenance is performed at specific points in time based on systems survival and failure history. However age is not the only determinant factor affecting the time-to-failure of a system. An older healthy system may have higher reliability than a newer, but deteriorated system. Degradation and health state of a system can play major roles in its time-to-failure and consequently the maintenance system cost. Condition-Based or predictive maintenance approach has been used by many analysts and practitioners in the last few decades to reduce the maintenance system cost. Depending on the type, and accuracy of the information available, different approaches have been proposed and investigated. This study compares the long-run average cost of various maintenance policies, fed different level of information accuracies, when the Proportional Hazard Models is used, and identifies the value of accurate condition monitoring information. The accuracy ranges from “no condition monitoring data available” to “perfect condition monitoring information” regarding health state. Furthermore, we compare the CBM strategies long-term average cost with that of Time-Based maintenance, and Run-to-failure strategies. Failure data is generated through simulation and long-run average costs of maintenance strategies are calculated for different range of the model parameters. This study offers a guideline for maintenance decision makers to select the suitable maintenance strategy.