Rolling element bearings are critical components in most rotary machines. When a bearing is damaged, it tends to generate higher vibration than a normal bearing does. Therefore, to prevent machine breakdown due to bearing damage, bearing vibration is normally monitored in critical industrial machines. However, during the early stages of bearing damage, the vibration signal is weak compared to those from other components such as gears which makes early detection very difficult in most cases. To overcome this problem, AE (Acoustic Emission) technology was evaluated for potential condition monitoring of rotary machines. Results presented in this paper show that AE can help detect bearing damage at an early stage.