Aimed at the existed uncertain factors affecting the performance of the servo system in the XY table driving by permanent magnet linear synchronous motor, such as the system lag, non-modeled dynamics, system parameters variation, friction, external load disturbance and so on, a kind of recursive neural network control method is proposed to control the operation state of the servo system, so as to achieve the robust precision tracking control purposes. Since the recursive neural network was a dynamic mapping structure, the uncertainty mentioned above could all be controlled effectively, which had two kinds of feedforward and feedback connection methods for network. At the same time recursive neuron had an internal feedback loop, the dynamic response of the system could be obtained without external delay feedback. The experimental simulation results show that, the designed control system has quite strong robustness and fast tracking performance, which can greatly reduce the contour error of the system, and improve the positioning accuracy of the system.