Aiming at the fact that the traditional fixture positioning can not meet requirements of complex curved blade damage in the process of aero-engine blade repair, an adaptive positioning method for blade repair on complex curved surface based on speckle measurement was proposed. Firstly, a visual measurement system was constructed. The three-dimensional configuration of the blade was achieved by matching the speckles on the blade surface and calculating the spatial coordinates of the reference measuring points and the data of the blade solid point cloud. Then, an adaptive positioning model was established, and the pose matrix of each positioning coordinate system was calculated to realize the adaptive positioning of curved surface blades under under-positioning clamping in machine tool coordinate system. Finally, an adaptive positioning experiment based on demage five-axis NC machining center was carried out to verify the effectiveness and practicability of the positioning method. The experimental results show that the deviation of translation vectors is less than 0.25 mm and the deviation of Euler angle of rotation matrix is less than 0.2°. The shape and position information of blades are extracted by one-time clamping and fixing using visual measurement to reduce the locating process, which can provide an experimental basis for the subsequent adaptive positioning and repair of complex surface blades.