Intelligent Transportation Systems have recently received great deal of attention and Vehicular networks and its applications represent a major part of ITS. Many vehicular network applications require accurate location information to improve their performance. Over the past years, many researchers worked on state prediction/estimation techniques in tracking, navigation applications for mobile ad hoc networks and wireless sensor networks. Yet, few were into the field of Vehicular networks. In this paper, We study five different movement prediction models and their efficiency and effectiveness for VANETs. We compare them using both real vehicle mobility traces of taxi cabs and generated mobility traces from SUMO.