Abstract: When the traditional sliding mode control ( SMC) method is applied to steer-by-wire ( SBW) system，it is necessary to obtain the upper bound value of system disturbance in advance，and the change of system disturbance will lead to poor stability of angle control． In order to improve the wheel angle tracking performance of SBW system，an adaptive neural network sliding mode control ( ＲBFSMC) method considering system disturbance was proposed． Firstly，the ＲBF neural network was used to estimate the system uncertainty and motor torque disturbance in real time． Secondly，the wheel angle controller was designed by combining ＲBF with SMC in order to improve the adaptability and stability of the angle control． The joint simulation results of MATLAB /Simulink and CarSim on SMC and ＲBFSMC show that，ＲBFSMC can better maintain 0° wheel angle and realize fast and stable tracking of dynamic wheel angle under the conditions of vehicle maintaining steering，continuous steering and single /double lane changing，and it has better corner response and tracking performance than SMC． The study indicates that ＲBFSMC has better robustness and stability than SMC．

Key words: steer-by-wire, radial basis function neural network, adaptivity, sliding mode control