The most used technology for ITS (Intelligent Transportation System) is VANET (Vehicular Ad-hoc NETworks) which is a subclass of MANET (Mobile Ad-hoc NETworks).VANET enables wireless communication between vehicles as well as RSU (Road Side Units), by using the standard 802.11p channels bandwidth to transmit all sort of information to each vehicle, within the range of the communication, providing passengers with plethora of safety, comfort and traffic monitoring applications. These networks are characterized by the predictable motion nature of vehicles which allows to predict the future locations of the nodes. They are also known for their varying topology through time, due to the changing number of nodes and the lane-constrained mobility patterns and their reduced power consumption requirements. However, the implementation, deployment and testing of VANET implies significant cost and requires a huge experimental platform. Therefore, computer simulation seems to be one of the best alternatives to test VANET communication protocols, before their deployment. However, the correct simulation of these networks requires to have access to a large number of heterogeneous mobility and traffic scenarios. Several research works try to solve these issues by creating mobility models that tend to accurately generate trace-files, that reflect the true nature of road traffic motion. This paper reviews these existing solutions and the need to develop them into intelligent tools, capable of faithfully reflecting human driving behavior in computer simulation scenarios designed to accurately test VANETs.