Abstract: A behavior prediction and control method in an intersection for intelligent vehicles was proposed. A finite state machine (FSM) model was established to predict the behaviors of other vehicles in the intersection. A hybrid state system for an intelligent vehicle was built and the FSM model was used to analyze the district states combined with the safety rules of intelligent vehicles and to control the vehicle under the safety condition. A co-simulation with PreScan and Simulink/Stateflow was conducted. The experiment results show that the intelligent vehicle can safely go through the intersection with other vehicles using the proposed approach.