ABSTRACT: Simulation of wheel and rail wear allows to predict long term profile evolution and thus, study the consequences of wheel damage in the dynamic behaviour of the vehicle, or study future maintenance requirements. Several models have been developed which try to solve the wear issue by relating the energy dissipated in the wheel-rail contact to the worn out material, from which two can be highlighted (Tg/A and Archard) which have significant differences on contact level. Even though, the prediction of long term wheel profile evolution has been validated with these two models, which means that for regular applications they seem to have an equivalent behaviour. In this work similarities and differences between the long term wear prediction methodologies are analysed, discussing their actual limitations. Then, these differences are exploited in specific operational cases to compare their wear prediction performance.