In order to realize the intelligent shift of robot driver, a shift control method of vehicle robot driver based on fuzzy neural network (FNN) was proposed. The displacement of throttle pedal for robot driver, speed and acceleration of test vehicle were used as the input of FNN model, and shift was used as the output of FNN model. The number of membership functions was three, and the type of membership functions was gbellmf (generalized bell membership function). The hybrid learning algorithm that combined back propagation algorithm with least square method was applied to train the model. Simulation results demonstrated that the results of simulation shift for robot driver using FNN control had a good consistency with the results of experimental shift. Furthermore, the proposed method could realize the gear-selection of robot driver correctly with the changes of operation environment.