The main purpose of condition monitoring in gear systems is to facilitate condition-based maintenance (CBM) by detecting the degradation of components, e.g. gear teeth, before the occurrence of a failure which could result in a malfunction of the whole gearing system and a reduction of the system availability. The role of vibration-based condition monitoring is to detect any change in the dynamic response related to changes in the structural integrity or the excitation forces. The early detection of a degraded gear allows a properly scheduled shutdown, which could prevent catastrophic failures with cascading effects and consequently result in safer operation and a reduced maintenance cost. The current article reviews the methods applied for gear tooth fault detection using vibration analysis, with an emphasis on tooth fault modelling and dynamic simulation. The article concludes with a brief discussion of the methods and models that have been used during the past few decades.