The invention discloses a three-dimensional reconstruction method for a light field of a high-reflection surface structure based on point cloud self-adaptive repair. The method comprises the following steps: firstly, building a structured light field three-dimensional reconstruction system, and extracting a multi-view light field sub-aperture image; then, calibrating the reconstruction systemcalibrating , projecting the sine stripe sequence projecting to a reference plane, moving the reference planemoving, recording phase changes recordingon the reference plane under each moving distance through a focusing type light field camera, and subtracting the phase changes from phase values of pixel points at an initial position to obtain a phase difference-depth-of-field (moving distance) mapping relation; reconstructing to obtain a multi-view three-dimensional point cloud through a sub-aperture image extraction algorithm and a mapping relation; and finally, multi-view point cloud data obtained through reconstruction having certain complementarity. A data cavity part in the target point cloud can be iteratively repaired by the same position of other view point clouds, so that complete three-dimensional reconstruction of a high-reflective surface object is realized. In addition, aiming at the reconstruction result, the invention provides a corresponding evaluation method.