Abstract: Objective To support quality control and process optimization in multi-layer multi-pass welding, this study aims to employ line laser sensing detect weld geometric features across the pre-weld, interpass and post-weld stages. Methods Surface point cloud data of the weld are acquired. By combining slope analysis and inflection point detection, three core modules are constructed: groove detection, interpass detection, and post-weld detection. In groove feature detection, key points are identified based on slope analysis. Inflection point detection extracts the deposited weld bead and its geometric features. Reinforcement height and cross-sectional area are computed after welding. Results In multi-layer multi-pass welding tests, the method effectively identifies key geometric features including groove depth, weld width, reinforcement height and cross-sectional area. The method exhibits high detection accuracy under the tested conditions. The positional consistency of the detected features was verified by comparing them with the macroscopic cross-sectional morphology. Conclusion The adopted line laser geometric feature detection method provides an integrated workflow for the pre-weld, interpass and post-weld stages, demonstrating practical engineering value and offering technical support for quality control, process optimization and automation in multi-layer multi-pass welding.