To effectively improve the accuracy of weld seam trajectory tracking results for H-shaped steel welding in intelligent construction， a weld seam trajectory tracking method for H-shaped steel welding in intelligent construction based on machine vision is proposed. The surface images of H-shaped steel in intelligent construction are collected by a machine vision system and denoised by combining non-local mean filtering and bilateral filtering. Texture analysis is performed on the surface images of H-shaped steel in intelligent construction to extract image texture features. At the same time， the weld area is determined by image segmentation. The Shi-Tomasi algorithm is introduced for corner detection. The inflection point information corresponding to weld groove information is obtained by fitting a straight line to find the intersection point and scanning line by line. The weld center is determined to realize weld seam trajectory tracking for H-shaped steel welding in intelligent construction.Through actual test and analysis， it is proved that the proposed method can accurately track the weld seam trajectory and has strong real-time performance， and it can effectively promote the development of intelligent construction.