The work aims to control the liquid level of the liquid storage tank in the beer filling machine by setting the optimum PID control parameters to ensure that the PID controller can meet the control requirements in beer production. The particle swarm algorithm was improved by Rosenbrock search method and individual disturbance, and the improved algorithm was used to set the PID parameters, and finally the PID controller with setting parameters was used to control the liquid level. The simulation experiment was carried out based on Matlab and the proposed algorithm was compared with the particle swarm algorithm. Through Matlab simulation verification, the PID parameters set by the improved particle swarm algorithm had a tracking characteristics adjustment time of 16.18 s, an overshoot amount of 10.20%, and an IAE performance index of about 6.09. For the particle swarm algorithm, the adjustment time of tracking characteristics was 27.72 s, the overshoot was 26.90% and the IAE performance index was about 7.23. Compared with the original particle swarm algorithm, 3 PID parameters set by the improved particle swarm algorithm in this paper have better comprehensive performance evaluation indicators when controlling the liquid level, and makes the system transition smoothly with less overshoot, fast response speed and fast adjustment time, so that the controller performance can meet the production requirements of beer filling machines.