Development history of maintenance Currently, maintenance methods that primarily practice in the industry are Preventive Maintenance, and vibration measurement is mainly used for rotating machine diagnosis. Preventive Maintenance is performed through trend management of vibration signals. When faults occur, it is possible to analyze faults by analyzing Bode Plot, FFT Spectrum, or Orbit Plot. However, the cause of the vibration signal can be various and complex and defects cannot be clearly detected by the vibration signal, so it is difficult to make an accurate diagnosis. In this paper, through the various critical features of the acoustic emission signal and vibration signal by representation, extraction, selection and trends were investigated for early detection of the possible failure of the rotating machine. The result of FFT analyzed for 7 times during 89 hours using the frequency analysis. It is very hard to detect early misalignment using the frequency analysis methods. However, the results of features analysis detected a fault growth in the rotating machine.