Abstract: The continuous escalation in operational speed and service mileage of Chinese rail transit vehicles has induced fretting wear in the wheel and axle department, emerging as a pivotal factor restricting the secure railway operation. Focusing on the multi-mode composite fretting wear within the interference fit between the wheel and axle, this study investigated the composite fretting wear characteristics of EA4T axle steel under cylinder/cylinder orthogonal contact, utilizing a self-developed composite fretting wear tester. The study explored fretting wear behaviors, accumulated dissipated energy, and wear mechanisms under composite fretting wear. These findings established a foundation for damage failure analysis of the wheel and axle fit part of rail transit vehicles. The system employed a horizontally mounted voice coil motor to propel tangential reciprocating motion. A tangential grating displacement encoder captured real-time tangential displacement, enabling closed-loop control. Similarly, a vertically mounted voice coil motor applied radial alternating load, and a radial force sensor collected real-time radial load, facilitating closed-loop control. The variation period for radial force and tangential displacement remained identical, ensuring that the radial force reached its maximum amplitude when the tangential displacement reached its peak. Three fundamental types of F_t-D-N curves were identified: quasi-trapezoid cycle, trapezoid and elliptic alternating cycle, and linear cycle. These corresponded to the slip regime (SR), mixed fretting regime (MFR), and partial slip regime (PSR), respectively. Wear scars exhibited a cometary appearance, with the side under a smaller radial load displaying a greater potential for relative slip, causing deeper wear marks. In the SR, the friction coefficient experienced a brief decrease after a rising phase, followed by a smooth increase until stabilization. In the MFR, the friction coefficient underwent a more prolonged decreasing phase, briefly stabilizes, and then continued to increase. The primary wear mechanisms of SR and MFR included fatigue spalling, abrasive wear, and oxidation wear. In the SR, the F_t-D-N curves assumed an asymmetrical quasi-trapezoid cycle, with the wear volume per cycle initially increasing before decreasing with the growing number of cycles. The area where chemical friction occurred also gradually expanded, leading to a rise in oxygen content. The object’s cross-section revealed significant delamination, indicating severe wear. The wear mechanism progressed from slight plastic deformation to fatigue spalling in the early stages, later manifesting as fatigue spalling, abrasive wear, and oxidation wear. In the MFR, wear was mitigated by slight ploughing and delamination at wear scars, with visible cracks extending into the subsurface. In the PSR, an adhesive region existed in the center of the contact zone, and a microslip region was observed at the edge of the contact zone. Worn pits were evident in the cross-section, with adhesive wear been the primary wear mechanism.