The work aims to propose an improved neighborhood based differential evolution algorithm and apply it to PID parameter optimization setting to solve the problems of difficulty, nonlinearity and hysteresis in beer liquid level control system, thus improving the work efficiency of filling machine and the beer quality. The differential evolution algorithm was improved to design a new mutation strategy and introduce the neighborhood search operation in the mutation link. According to the current population distribution, real-time adaptive allocation of the number of neighborhoods was carried out to enhance the global and local search ability of the algorithm. The proposed algorithm was compared with 2 basic differential evolution algorithms and 4 improved differential evolution algorithms and its performance was verified by 18 test functions. From the simulation results, compared with the basic differential evolution algorithm, the PID parameter setting by the improved differential evolution algorithm could reduce the adjustment time by 0.22 s, the rise time by 0.04 s, and the overshoot by 7.63%. Through the improved differential evolution algorithm, the PID parameter setting of the beer filling machine’s liquid level is optimized, which significantly improves the overshoot, rise time and steady-state error of the control system and realizes the stable control of the liquid level.