The condition monitoring of piston–cylinder-reciprocating machinery usually relies on vibration and acoustic sensors installed on the outer surface of cylinders. However, vibration and acoustic signals are susceptible to external interference from other accessories, and require an external power supply, which limits its widespread application. Herein, based on the lateral sliding mode of triboelectric nanogenerator (TENG), a novel reciprocating device with condition-monitoring and self-powering capabilities is proposed, called piston–cylinder triboelectric nanogenerator (PCTN). The effects of different factors, including mean piston speed, number of piston rings, materials of piston ring and cylinder, on the output characteristics of PCTN are investigated, respectively. Two typical fault cases, that is, piston ring missing and coil fracture, are investigated to verify the condition-monitoring capability of PCTN. Piston ring missing faults can be effectively identified based on the variations in peak and root-mean-square (RMS) values of short-circuit current (I sc). Coil fracture faults can be identified and located by analyzing changes in time–domain curve and time–-frequency spectrum of I sc. Herein, a theoretical and experimental basis for the widespread application of PCTN in reciprocating machinery is provided.