The advent of Industry 4.0 has significantly transformed the manufacturing sector by leveraging new digital technologies, automation of production processes, plant connectivity (IoT), and artificial intelligence (AI). In such a context, where the company is inclined to invest in digital transformation to become a smart factory, predictive maintenance is becoming increasingly relevant. Out of various monitoring techniques used in the industries for maintenance, vibration monitoring is more reliable in predicting the machinery health. We know that state-of-the-art techniques like machine learning are useful for general prediction problems. “Predictive maintenance of rolling element bearings using vibration analysis and machine learning techniques” is a proactive approach to maintenance over conventional techniques. While machine learning offers promising solutions, this research introduces a novel, highly efficient model for accurately predicting bearing failures. By doing so, it aims to predict when these bearings are likely to fail or require maintenance before a breakdown occurs. In this research paper, the vibration response of the rolling bearings to the defects on the outer race, inner race, and the rolling elements is obtained using an accelerometer sensor, data acquisition system, and then analyzed. It shows that every defect excites the system at its characteristic frequency. The location of the faults is indicated by the FFT spectrum. Defects are indicated at both bearings in a horizontal direction. The results reveal that the vibration-based monitoring method and developed machine learning model is successful in detecting the faults in the bearing for predictive maintenance of critical machiner