Abstract: For the problem caused by nonlinear friction in CNC machine at low velocity,a parallel hybrid control model based on neural network and PID algorithm is proposed.While the motor speed is higher than the switching speed,the PID algorithm is adopted,otherwise,the neural network controller is used.The neural network is a single neuron with five inputs,and uses Hebb learning algorithm.Analysis shows that the hybrid controller greatly decreases the variation of following error,and the machine runs more smoothly.Experiments are made with the programmable multi-axis controller(PMAC) to validate the performance of the hybrid controller.