In this paper, there is a control system which has a strict requirement on the system load and a limited installation space for the sensors. According to the conditions of this system, the scheme by only installing a group of angular position sensors on this system is proposed. Nevertheless, the tracking performance of the system is limited by the time delay of the CCD, the traditional feedback control methods cannot meet the requirements of the system tracking accuracy. The feedforward method can theoretically improve the tracking performance without affecting the dynamic performance of the system. Unfortunately, the feedforward control method requires accurate target state information, and the CCD can only detect the light-of-sight error of the target. Therefore, we propose recovering the target trajectory by combining LOS errors from the CCD and the platform angular position from the angular position sensors, and adding a position feedforward to the system based on a dual-closed loop of angular position sensors and CCD. In addition, there is bad noise in the target state information obtained by the fusion. Then, a Kalman filter is used in this paper to filter and estimate the target state information. Essentially, the Kalman filter is the feedforward controller of the system. In order to design a better feedforward controller, a frequency-domain analysis method for the time-domain filters is proposed. The methods proposed in this paper are verified trough simulations and experiments.