In view of the isolation problems caused by the seeker system internal disturbance torque, this paper analyzed the main factors influencing the disturbing torque. A model for tracking loop was made by the neural network, and iterative learning and training of the influencing factors was made to produce a similar disturbance torque to give feedforward compensation of tracking control loop. It reduced the interference of the system output disturbance torque, while increased the performance of the missile. Simulation results showed that the artificial neural network was effectively introduced into seeker tracking control circuit to solve the problem of isolation caused by disturbance torque, reduce the miss distance and increase the accuracy of missiles.