<p>Prediction of wheel wear for the railway vehicle can be used to estimate the wheel operation state and evaluate maintenance intervals. This paper proposed a wheel wear prediction method for the railway vehicle running in the complex line. The forecasting approach incorporates a wheel wear assessment framework that integrates the vehicle dynamics system, a wheel-rail rolling contact model in finite element, and a wear rate computation model. The track irregularity, curve line proportion, different rail profile are all considered in the method. The predicted results of wheel wear are compared with the measured results for a Chinese intercity trains. Simulated and mesured results show the wear coefficient map measured by Jendel is not accurate to predict the wheel wear of the studied CRH6A trains. The wear coefficient correction factor equaling to 2.5 is proposed for the studied vehicle. The wear of flange in simulated and predicted results both show three periods: Initial fast wear period, middle non-wear period, later serious wear period. The changes of flange wear rate are mainly caused by the changes of wheel and rail profiles. The predicted results demonstrate excellent agreement with the measured data in terms of both tread wear and flange wear parameters. So the proposed prediction method could predict the wear of wheel running in the complex line containing many small curves accurately.</p>