Wind turbines are devices that can be used to change kinetic energy from wind to electrical energy with the help of generators. One component that is often damaged is bearing. Damaged bearings will affect the performance of the wind turbine, so that the energy produced by the generator will decrease. The purpose of this study is to detect damage to spherical elements from ball bearings based on vibration signals on horizontal axis wind turbines with analysis of time domain statistics using themethod Principal Component Analysis (PCA). This study used bearings with normal conditions and bearings that are intentionally damaged on the ball elements with a depth of 2 mm and a width of 0.7 mm. Bearing damage detection is carried out using 7 time domain statistical parameters and Principal Component Analysis (PCA). Data retrieval is done using a motor as a substitute for wind with a rotating speed of the shaft at a bearing of 1200 RPM. The bearings used are Self Aligning Double Row, Brand TAM, Series 1208K. Extraction results of 7 statistical parameters 4 statistical parameters can distinguish the conditions of both bearings and the other 3 statistical parameters have not been able to distinguish the conditions of the two bearings. The Principal Component Analysis (PCA) method is applied to utilize important information on these statistical parameters. The results of the study showed that thedomain-based method PCA proposed was successful in reducing data from 7 statistical parameters so as to provide new information (4 PCs) with maximum variance. The data information contained using 2 PCs was 86.421% while with 3 PCs it was 94.463% variance. Themethod is Principal Component Analysis (PCA) able to clearly identify and classify the differences between normal bearing conditions and the condition of damaged ball elements in horizontal axis windmills.