The condition monitoring system of gearboxes usually employs accelerometers fixed on the gear housing to diagnose the gear damage. Such diagnostic system is essential in scheduling the maintenance of drive-Trains. However, the successful application of vibration based diagnostic techniques for planetary gearboxes is still challenging because of the complexities of the sideband behaviors in the measured vibration signal, which is a result of the revolution of the planets around the sun gear. Moreover, the performance of diagnostic techniques may further deteriorate since the measured vibration signals are often contaminated by the noise. Therefore, it is essential to develop a diagnostic scheme for extracting the fault signatures in planetary gear-sets. This paper begins with the summary of an analytical model of the planetary gear-sets developed by the authors that describes the characteristic behaviors of the fault introduced sidebands of the vibration spectrum picked by the fixed vibration sensor. This analytical model, which predicts the frequency components of planetary gear-sets, is easy to implement because the only prerequisite information is the basic geometry of the planetary gearbox. Afterwards, a novel vibration based diagnostic scheme for planetary gear-sets is developed. Its performance is then validated using simulated vibration signals generated from a dynamic model of the planetary gearbox. The proposed diagnostic scheme integrates the analytical vibration model, a de-noising algorithm and a frequency-domain indicator into one synergistic system for the early detection of damaged gear teeth in planetary gear-sets.