Energy system models are often used to assess the potential role of hydrogen and electric powertrains for reducing transport CO2 emissions in the future. In this paper, we review how different energy system models have represented both vehicles and fuel infrastructure in the past and we provide guidelines for their representation in the future. In particular, we identify three key modelling decisions: the degree of car market segmentation, the imposition of market share constraints and the use of lumpy investments to represent infrastructure. We examine each of these decisions in a case study using the UK MARKAL model. While disaggregating the car market principally affects only the transition rate to the optimum mix of technologies, market share constraints can greatly change the optimum mix so should be chosen carefully. In contrast, modelling infrastructure using lumpy investments has little impact on the model results. We identify the development of new methodologies to represent the impact of behavioural change on transport demand as a key challenge for improving energy system models in the future. Copyright (C) 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.