Abstract Bearing fault diagnosis is of great significance for ensuring the safe operation of equipment and promoting the development of intelligent manufacturing. Based on the energy harvesting technology of triboelectric nanogenerators (TENG), this paper designs and fabricates a TENG-based smart bearing with self-powered sensing function, and proposes a bearing fault diagnosis method based on the data fusion of vibration signals and triboelectric signals. The triboelectric signals are generated by a TENG sensor unit, which is integrated into the bearing structure and is more sensitive to minor changes in the bearing’s operating state. These triboelectric signals exhibit differentiated characteristics from vibration signals in terms of perceptual dimensions and physical mechanisms. Through the fusion of the two types of signals, their complementary advantages can be effectively exerted to obtain more comprehensive bearing condition information. Experimental results show that the proposed fusion method demonstrates excellent diagnostic performance under various noise conditions, which is significantly higher than the methods using a single type of signal. In turn, this study provides a novel data fusion diagnostic approach for bearing fault diagnosis.