A visual sensor is used to measure the vibration of a piezoelectric flexible cantilever plate. To decrease the image processing time and satisfy the requirement of real-time application, the searching range is limited by selecting region of interest images and using a tracking technique. After image processing, including median filter, image segmentation for object recognition, least squares technique, the vibrations of bending and torsional modes for flexible plate can be obtained. The visual measured signals are used as feedback to suppress the excited vibrations. The conventional proportional plus derivative (PD) control and finite-time control algorithms are designed and implemented. The experiments are conducted. The decoupling of measurement of the bending and torsional vibrations is realized by the visual sensor accompany with applying the image processing methods. Experimental results demonstrate that the designed control algorithm can suppress the low-frequency vibration effectively. Furthermore, the finite-time controller can damp out the small amplitude vibration much better than that of the PD control.