It is the key issue of tracking travel that intelligent vehicle travels fast, smoothly and accurately in curved path. An intelligent vehicle control scheme based on CCD camera path identification is presented in this paper, and the intelligent vehicle can achieve full closedloop control at different stages of the curved path. During the travel process, the intelligent vehicle recognizes the straight and curved paths according to the curvature of the navigation line. Different control methods are adopted in the straight and curved sections of the path to realize the advanced deceleration for the curved section and the accelerated passing through for the straight section, and the full closed loop control is adopted in both the straight and curved sections. The visual system extracts the navigation line in the path with the CCD camera and calculates the curvature of the navigation line as well as the parallel deviation value and course angle deviation of the intelligent vehicle relative to the navigation line. The main control system selects the control strategy according to the curvature of the navigation line, calculates the steering angle and travel speed according to the parallel deviation value and course angle deviation, and completes steering and speed adjustment through the actuator. The control results are fed back to the main control system through the visual system and speed measurement device to achieve the full closed loop control. An experiment system was set up in the background of the 15th national college student robot competition, through multiple debugging and experiments, the maximum travel speed of the intelligent vehicle on the track reaches to 35 m/s, the average speed in the whole course is about 28 m/s, and the actually measured maximum deviation distance is 35 mm, which is within the allowable error range. The actual experiment system verifies the feasibility of the staged full closedloop control method of intelligent vehicle.