During the measurement process of Airborne LiDAR Bathymetry (ALB), there are problems such as difficulties in setting control points and residual calibration errors. At the same time, due to the inconsistent accuracy of underwater measurement points, there is an elevation inconsistency phenomenon between ALB survey strips.First, the overlapping area between strips is extracted based on the eight-neighborhood method to limit the point-surface matching range. Then, by constructing a Triangulated Irregular Network (TIN) and matching points of adjacent strips to determine approximate corresponding points, the relationship between strips is established. The Random Sample Consensus (RANSAC) algorithm is used to optimize the matching, a regional network strip adjustment model is constructed, and the optimal transformation matrix of the strips is solved. Finally, a polynomial surface is used to represent the complex terrain, and the correction values of each point are calculated and corrected according to the point-surface matching distance and the least-squares solution of the polynomial coefficients.To verify the effectiveness of the proposed method, experiments were carried out using data collected by the ALB system Mapper 20KU, and the data accuracy before and after adjustment was evaluated based on land RTK points and shipborne single-beam bathymetry points. After the ALB strip adjustment, the land and underwater measurement deviations decreased by 8.8 cm and 7.5 cm respectively, and the processed data bathymetry accuracy was 24.0 cm.This paper takes into account the data characteristics of the ALB system and the limitations of measurement operations. Considering that the ALB point cloud data in coastal zones lacks obvious features, and the strip data is approximately planar with sparse point clouds, the eight-neighborhood overlap region extraction method is introduced to improve the data matching efficiency and avoid local optimal solutions caused by iteration. In view of the elevation discrepancies between strips that affect the representation of real terrain data, the regional network adjustment with point-to-surface matching, combined with the RANSAC iterative method, effectively improves the internal consistency accuracy of ALB. To address the uncertainty of external consistency accuracy caused by the lack of control point constraints in regional network adjustment, the intersection area of the adjusted survey lines and inspection lines is used as a control, and the Bursa model is employed for correction. Considering the inconsistent accuracy of ALB point clouds above and below the water, a nonlinear adjustment model is utilized to weaken the distortion within the strips.