Rolling-element bearing (REB) Failure is one of the most common failure in rotary machines in industry. Therefore, this concern draws attention to the concept of diagnostic and the prediction of REB failure. The efficacy of vibration analysis in this area has led many researchers to focus on this topic in recent decades. Defective bearings such as fatigue, surface roughness, and corrosion can lead to disastrous accidents. Therefore, providing an algorithm for the detection and classification of bearing defects is valuable and useful. The prediction of the bearing's lifetime is important to ensure the planned production in an industrial unit. In this paper, different models for performance degradation of mechanical parts and signal processing techniques are presented for fault detection in ball bearings. Then, the failure modes of bearings and their relations to the final fault using the vibration signals are discussed. Thereafter, by examining the experimental data of the PRONOSTIA test and using the root mean square of signals, the degradation performance of the bearings and the corresponding failure patterns are obtained. In addition, using the results of frequency analysis via Envelope method, the initiated/existed defects on the bearings of each test are detected. In the application of this method, selection of appropriate frequency band for filtering the vibration signal is important. By analyzing the results, it can be concluded that any ball bearing failure pattern does not necessarily refer to a specific defect. Also, the proper area for filtering a signal in envelope analysis is not necessarily a resonance region, but it should also check several areas to select the best area.