Abrasive wear is one of the most critical degradation mechanisms in tribological systems, leading to material loss, reduced efficiency, and premature failure of components. Despite its importance, predicting the initiation and progression of abrasive wear remains challenging because the complex nature of lubrication conditions. In this study, continuous pin-on-disk experiments using bearing steels were conducted to investigate predictive indicators of abrasive wear. During the first 2,000 s, the contact was lubricated with a base oil to establish stable frictional conditions. At 2,000 s, an abrasive slurry was introduced into the contact without interrupting rotation, simulating the sudden intrusion of abrasive particles. Subsequently, the evolution of friction before and after the onset of abrasive wear was analyzed from three complementary perspectives, demonstrating the potential for predictive assessment. First, the spatial distribution of friction on the disk under both lubricated and abrasive wear conditions exhibited similar patterns despite the large difference in magnitude, suggesting that friction measured in the lubricated regime already reflects the inherent spatial structure of the contact and can therefore indicate where high friction and wear are likely to occur. Second, wavelet analyses of friction signal successfully revealed transient features in the frequency domain that are associated with the initiation of abrasive wear, with spectral responses appearing before the increase in friction. Third, precontact measurable parameters, namely, disk surface topography and disk tilting induced oscillations, were found to be effective indicators for predicting high friction regions. These three approaches provide a framework for forecasting, detecting, and understanding abrasive wear, thereby contributing to the development of intelligent predictive maintenance strategies in tribological systems.