In order to obtain driving characteristics of the piezoelectric actuator, a sub-pixel displacement measurement method based on micro-vision is proposed. The initial position is solved by the Harris corner location method. The initial template matching region is obtained by corner matching to critena raster image sequence. The sub-pixel displacement measurement method of duality Lagrange interpolation and nonlinear Newton-Raphson iterated algorithm are proposed, which can effectively reduce the time-consuming for sub-pixel matching related calculations to ensure the good measurement accuracy. A modified sub-pixel image block matching method is used to measure the actual driving characteristics of piezoelectric actuators, and the measurement result is compared with the real displacement and PI model. Experimental results show that the algorithm of displacement measurement exhibits high accuracy and good stability. Piezoelectric characteristic curves obtained agree with the results of physical analysis and can satisfy requirements of high-precision measurements of nanomanipulation.