Facing the problem of low efficiency and unstable quality in the traditional hand finishing tricot lace cutting field, a novel tricot lace real- time recognition method based on feature recognition is proposed. The method works by adaptive thresholding, line detection, image segmentation and reducing the fake contours, which can recognize accurately the profile features and locations of tricot laces. It has high stability even in complex factory lighting environment. It can satisfy the real-time request and promote computational efficiency by improving the computational efficiency of the method. The automatic vision cutting system based on the proposedmethod can highly improve production efficiency and quality.