A visual servo method based on image trajectory planning and sequential control was proposed for uncalibrated eye-in-hand robot system. First the trajectory planning method in image space and task space was presented. Second the Kalman filter was adopted to estimate the relative pose of the target to camera, and combined with camera measurement, the adaption of Jacobian matrix was realized. A visual servo controller with image feedback was designed to realize the visual tracking of target. At the end of this paper, the algorithm was verified by MATLAB simulations and experimental researches, and compared with classical PBVS and IBVS method. The results verify the effectiveness of the algorithm.