This paper mainly studies the speed and robustness of the single-axis control system of the gyro stabilized platform. The sliding-mode variable structure is combined with neural network for controlling the gyro stabilized platform. Aiming at the chattering problem of traditional sliding mode variable structure method due to inaccurate model estimation, a neural network sliding mode variable structure control method is proposed using RBF neural network to approximate the unknown part of the model. Experiments show that using this method can effectively improve the systems speed and robustness.For the uniaxial control system of the gyro stabilized platform, the method can implement the angle tracking in 0.3 s, and the speed tracking in 0.6 s;the maximum error of angle tracking is 0.007 3, and the error converges to a neighbourhood of zero in 1 s.