To eliminate the problem that the quantization and scale factors in the fuzzy control of the constant force output device are fixed values, which results in the poor adaptive ability of the force control system, an improved particle swarm algorithm was proposed to continuously conduct iteration to select the optimal quantization and scale factors for the purpose of optimizing fuzzy PID control. The constant force output device was simulated and the transfer function was determined, the fuzzy PID control system and the fuzzy PID control system optimized by the improved particle swarm algorithm were established, and the two control systems were compared and analyzed via MATLAB. An experimental platform based on the six-degree-of-freedom TA6-R10 manipulator was built to verify the output force of the constant force output device. The results show that the fuzzy PID control system optimized by the improved particle swarm algorithm reaches the target value faster than the fuzzy control PID system, and converges faster, the overall overshoot is 8.91%, which becomes stable in 1.3 s. By using the fuzzy PID control optimized by the improved particle swarm algorithm, the force control of the constant force output device can be optimized, which has a faster dynamic response and a better force following effect.