The invention relates to an edge point extraction method and device based on a line structure laser point cloud, and the method comprises the following steps: 1, constructing a window according to theobtained coordinates of the line structure laser point cloud, and extracting a local point cloud P; step 2, determining a rotation angle theta in P; 3, integrally performing linear transformation onthe normalized laser point cloud coordinates according to the rotation angle theta, solving the maximum value and the minimum value of Y coordinates in the transformed point cloud, and judging to obtain candidate edge feature points k1; 4, sliding the window, continuously extracting the local point cloud from the second point, and extracting the candidate feature point k2; and step 5, judging whether the feature points k2 and k1 extracted by the two adjacent windows are the same point or not. According to the windowing extraction local point cloud processing provided by the invention, the number of point clouds participating in operation is greatly reduced, so that the calculation efficiency and the extraction accuracy are improved. Firstly, normalization processing is carried out on localcoordinates to ensure the influence of different coordinate values on the local coordinates in the point cloud extraction process.