Abstract: To address the limitations of traditional point cloud registration methods, such as strong dependence on initial poses, susceptibility to local optima, and constrained registration accuracy, a point cloud registration approach integrating MSAC and GICP is proposed. Firstly, the M-estimator sample consensus (MSAC) algorithm performs coarse registration to effectively remove outlier interference and obtain a preliminary transformation matrix. Subsequently, the generalized iterative closest point (GICP) algorithm is employed for fine registration. By integrating point cloud normals and covariance information, it optimized point pair matching, reduced registration time and improved accuracy, while enhancing robustness. Finally, registration experiments are conducted on post-scanning tee pipe point cloud data. Experimental results demonstrate that this method achieves high-precision registration under various orientation conditions. Compared to traditional ICP methods, the MSAC-GICP algorithm demonstrates significant advantages in error control and stability. When compared to the FPFH-RANSAC and RANSAC-GICP algorithms, it achieves higher registration accuracy, reduced registration time, and enhanced registration stability, proving the feasibility of this registration algorithm.