Aimed at the problem of vibration measurement and suppression of flexible structure, a visionbased vibration measurement and feedback control method of flexible structure is proposed. In order to ensure real time of the closed-lOop system, it is necessary to solve the problem of excessive processing time needed for mass data computing. A motion-tracking method is used to determine region of interest ( ROI ) of single frame to reduce the amount of processed data. Also, a fast median filter method is used to improve the speed of image processing. Roberts operator can improve processing precision of the beam’s tip edge detection, as compared with the other two differential detecting operators. Extremum and Hough transformation methods are used to extract and correct the edge information, respectively. Finally, theoretical analysis and experimental research of the vision-based vibration control are carried out. The results demonstrate that the adopted image processing methods meet the requirement of precision and real time, and also verify the feasibility of the vision-based vibration control scheme.