In an industry, fans are used in a variety of needs supporting the production process, namely the HVAC system, the combustion process, the cooling process, and air dryers which have quite an important role. One component that has an important role in the fan is the bearing. Bearing damage to the fan will significantly reduce the performance of the fan, even to the point where it can cause industrial fans to stop working. One method to monitor the condition of these bearings is a vibration-based method using spectrum analysis. However, spectrum analysis often experiences constraints in the analysis process caused by the large number of harmonics and sideband complexes that can be caused by several components present in the industrial fan. Therefore, cepstrum analysis will be examined in this study because the analysis can sort out the harmonic family which is complex of several components so that it can analyze bearing damage more easily. This study uses an industrial fan prototype with a transmission of one pair of gears and pulleys. The bearings used are ASB 6209 single row ball type bearings in normal conditions and inner race defects. An accelerometer sensor that is used to record vibration signals is placed above the bearing housing. As for the data acquisition process using the DAQ module from National Instrument. Recording is done for 10 seconds in one data file, a data set has 20 recorded data files. The results of this study indicate that the frequency spectrum analysis of internal path defects (BPFI) appears at a frequency of 54.69 Hz, but the peak is covered with high amplitude from other frequencies making it difficult to detect. Whereas the cepstrum analysis shows a peak at the 0.0201 s frequency which shows that the BPFI quefrency is quite clear and easily observable because cepstrum can sort the harmonic families on the spectrum graph.