A classifier plays a crucial role in the cement industry. It is in charge of separating coal that has been smoothened out and is ready to be burned although the coal is still rough after going through the grinding process. It takes a long time to burn coal that is not perfectly processed with a classifier. Therefore, it will reduce the amount of cement production, and the factories will release more energy. The closed arrangement and the number of components in the unit classifier requires a sophisticated method to detect damage that occurs early. Vibration analysis is a method that has been effectively employed in detecting the initial damage that occurs to the engine, especially the classifier. This study was aimed at detecting the location of the damage occurring in the classifier by using a vibration signal analysis and by measuring the magnitude of vibration and presenting it to the frequency domain (spectrum) form using Fast Fourier Transform. Engine condition assessment referred to ISO 10816-3 standard in velocity and displacement modes. Based on data spectrum analysis, the dominant damage laid in the unbalanced rotor. Spectrum characteristics of the damage appeared to be in the spectrum line worth 438.01 μm at a 3.5-Hz frequency (1X) radially. This analysis proved to be supported by the decrease in vibration value to 18.65 μm after balancing the Classifier rotor.