The work aims to reduce the impact of the working production environment on the weighing accuracy of the intelligent combination scale and increase the anti-jamming performance of the electromagnetic vibration system to the environment. The RBF neural network PID control algorithm was used to improve the electromagnetic vibration system in the production process. And the square of the momentum factor was added to the basic RBF neural network PID control algorithm. The empirical accumulation of parameter changes was considered to reduce the parameter adjustment. Compared with the basic RBF neural network PID control algorithm, the improved algorithm had faster convergence speed and better fitting accuracy. When the simulation length increased, the target function could still be well approximated. The improved algorithm makes the amplitude and vibration frequency of the electromagnetic vibration machine more stable. It can reduce the interference of noise in the environment.