Abstract: A new method was developed to detect the micro weld gap (0-0.1 mm) during laser welding,based on Faraday magneto-optical effect imaging. During laser butt-welding of carbon steel,a magnetic excitation device was used to magnetize the weldment. It was found that the magnetic field distribution in the weld zone was different from other zones. A magneto-optical sensor was placed upon the magnetized weldment to obtain the magneto-optical images of the weld. Different magnetic field intensities would lead to rotation of the polarized light of magnetooptical sensor in different directions,and the magneto-images consisting of the joint position feature could be captured. By using image-processing methods, such as median filtering, grayscale transform,threshold segmentation and morphology,the micro weld gap center could be extracted rapidly from a magnetooptical image. The experimental results show that the magnetooptical images could effectively and accurately reflect the position of micro weld gap. The proposed method provides a novel approach for automatic identification and tracking of the micro weld gap during laser welding.