Bearings are the most important components and are widely used on almost rotating machines. They play an important role in supporting and reducing the friction forces of the rotating shaft when working. Therefore, if there are any faults in the bearing it can cause loss of production and equipment as well as creating an unsafe working environment for people. Early detection of damage on the bearing helps to minimize the risk of major damage to the equipment, saving costs of machine downtime and repair. However, the bearing after replacement will quickly show similar damage if there are no measures to thoroughly resolve the cause of the damage. The location of the damage on the bearing is closely related to the cause of it. Therefore, the location of the damage is one of the useful pieces of information that helps the maintenance person find the cause and increase the chance of fixing the damage at the root. This is also the core strategy of proactive maintenance. In the current industrial 4.0 stage, artificial intelligence or AI is widely used as a tool to support problem solving. This study proposes the use of visualized scalogram of vibration signals, post processing with deep learning to identify the type of existing failure on rolling bearings. Time-based vibration signals of rolling bearing damage types are collected on a test rig. The vibration signals are then processed, and algorithms are applied on Matlab. The results of damage identification through vibration signals demonstrate the effectiveness of this method.