The work aims to solve the problem of low precision and slow speed of the traditional essential oil filling machine. A CMAC-PID liquid micro-filling self-tuning control system based on Kalman filter was designed. Firstly, Kalman filter was used to suppress the noise and disturbance signals. Then the CMAC-PID controller with the function to supervise learning was adopted to provide feedback control and keep the system stable. Combined with the system model, simulation verification was carried out through Matlab. The simulation results showed that, by using the CMAC-PID control method with Kalman filter, the system was characterized by fast response, small overshoot and strong anti-interference ability. The method can ensure the filling speed required by the system while ensuring the filling accuracy.