Industrial fan is a machine that is needed by an industry as another engine cooling device. One of the component that is often damaged in a fan is a bearing. The methods that are commonly used to monitor the condition of a bearing is a vibration-based method. The Time Synchronous Averaging (TSA) method is one of the vibration based method. TSA method is a method that is effective enough to reduce noise in vibration signals so that the peak amplitude of defect frequency on a bearing can be clearly seen. The TSA method was applied to ball bearing axial fan testing in this research. The purpose of this research was to reduced ball bearing signal from noise in industrial fan using the TSA method and detect outer race bearing defects and inner race defects in using the spectrum. Three types of ball bearings with different conditions, namely normal bearings, outer track defects, and deep trajectory defect in axial fan were tested. Tests are performed alternately with each data recording as many as 30 files. The data that has been recorded is then did a time domain plot and frequency domain plot in software matlab. The result obtained are then will be compared with the time domain plot and frequency domain plot at the signal preprocessing technique using TSA. The spectrum results obtained using the TSA method have better signal results than the signals obtained without using the TSA method. This is because the spectrum that TSA has done has less noise because it has been reduced. Noise that has been reduced using the TSA method causes the peak amplitude at the defect frequency of the ball bearing the inner trajectory and the outer trajectory seen clearly.