A Falex machine was used to perform the pin-on-disk wear tests. Two kinds of steel which included medium carbon steel and die steel were used. When sliding contact was under sufficiently high pressures and speeds, the pin wear became severe. By observing the edge of the contact interface during a test, we found that thin layers of debris were extruded out from the contact in the direction perpendicular to that of sliding. At the same time, flakes of debris were extruded from the contact in the sliding direction. Some of them piled up at the periphery of the pin, the rest formed sparks. Such phenomena are different from those resulting from oxidation, adhesive and abrasive wear mechanisms. The mechanism of this kind is now termed as extrusion wear. This study investigated extrusion wear mechanisms and transition of wear mechanisms of steel by micro-analyzing the wear surfaces and using mechanics of plasticity to evaluate and compare the real values and theoretical values of pin wear. Micro-analysis of medium carbon steel/ die steel pin shows that many extruded laminated layers were around the rim of the pins in the extrusion process. When these laminated layers accumulated to a bigger volume, they functioned as cooling fins. The contact temperature began to decrease and approached equilibrium. In this scenario, wear rate also decreased. However, the nominal contact temperature was maintained at over 600℃/700℃ in an extrusion wear regime. If the nominal contact temperature decreased to <550 °C, the wear regime changed to an oxidation one, and the extrusion process stopped immediately. In an extrusion wear condition, micro-analysis shows that layers of material were extruded continuously due to softening of the contact surface of the pin. Under such a condition, oxide films of negligible thickness formed during cooling in air after the wear test. The microstructure immediately close to the contact surface of medium carbon steel pin was fine pearlite, which was transformed from austenite at the end of the wear test. Below the fine pearlite zone, a coarse ferrite structure existed that resulted from martensite tempering (for hardened pins) or grain growth at a recrystallization temperature (for as-fabricated pins) during the sliding process. The microstructure close to the contact surface of a die steel pin was spheroidized carbide and main alloy carbide dispersed in untempered martensite base. Additionally, there were carbides dispersed in the tempered martensite base below untempered martensite. The hardening treatment of pins may influence friction and contact temperatures during the initial extrusion wear stage. When the contact reached a stable extrusion wear period, the real contact temperature usually exceeded 900 °C. In such conditions, the initial hardness of the pins did not affect the hardness of the contact surfaces in stable extrusion wear when running conditions were the same. Additionally, as the nominal contact temperature of the disc was lower than that of the pin, the sliding shear force or frictional force was mainly determined with the pin hardness. Therefore, the contact temperatures and friction coefficients for different heat-treated pins were almost the same. The real values or theoretical values of pin wear increase with normal load and sliding speed increasing, such a condition is consistent with the trend of extrusion wear. The results show that the real values of wear are of 1.7 to 5.5 times greater than the theoretical values. This inconsistency should be caused by some facts that were neglected in calculation. If the influent factors are modified appropriately, it will indicate that the plastic analysis is a feasible method to analyze the behavior of extrusion wear.