An artificial neural network methodology was applied to variable geometry turbine (VGT) modelling in order to enable representation of the VGT characteristics for any blade (nozzle) position. The VGT model was integrated with the diesel engine system, which was in turn linked to the driveline and the vehicle dynamics module to form a complete, high-fidelity vehicle simulation. Two VGT strategies were developed and analysed. A performance-oriented control strategy was aimed at increasing low speed boost and improving transient response. An EGR-enabling control strategy was aimed at providing the necessary pressure difference between the exhaust and intake manifold for exhaust gas recirculation (EGR) and emission control during normal driving. The effectiveness of the proposed strategies was evaluated for different driving scenarios for a Class VI truck, e.g. hard acceleration from stand-still and Federal Highway and Urban Driving Schedules. It was shown that VGT can virtually eliminate turbocharger lag. Furthermore, VGT is able to provide very accurate control of EGR level during highway driving, and a satisfactory level of EGR control during urban driving.