Considering the problem that the contact measurement influenced structure characteristics and the PID control parameter setting were not ideal in the vibration active control system, the machine vision technology was used to measure the vibration of the structure, and the vibration control was carried out with the PID optimized by artificial fish swarm algorithm. Rigid and flexible double-joint mechanical arm was selected as the research object. Firstly, the experimental platform was set up and the orthogonal test was designed to explore the vibration situation of rigid-flexible coupling mechanical arm and determine the motor setting parameters for subsequent vibration control; Secondly, CCD camera was used to capture the vibration image of the marker at the end of the mechanical arm, and the vibration displacement was obtained by processing and was used as the input of the control system. Finally, the artificial fish swarm algorithm was selected to optimize the PID parameters. The output signal of the controller was converted into control voltage by the output card. The control voltage was amplified by the power amplifier and drived the piezoelectric actuator to realize vibration control. The experimental results show that the average control effect of the optimized PID algorithm can reach 57.54% compared with the traditional PID control effect of 44.06%, which verifies the feasibility and superiority of the optimized PID control system based on visual vibration measurement.