A quantitative method for evaluating seam pucker in garment manufacturing is devel oped from three-dimensional image analysis using artificial intelligence. A laser scanning system created with a specially designed laser sensor measures the surface profile of seam pucker. Since the conventional AATCC rating system is not sufficient to quantitatively evaluate seam pucker, a new quantitative evaluating method is developed based on the geometric model of seam pucker. For shape descriptions of seam pucker, five shape parameters are proposed considering the surface of a seamed fabric puckered with wavy shapes. The shape of seam pucker can be defined as three-dimensional waves generated on the seam line and propagated to the edge line. The five shape parameters are the number of wave generating points, the wave amplitudes, and the wavelengths on the line next to the seam and on the edge line. A simulator produces puckered shapes when the five shape parameters are specified. For the simulator, puckered waves are assumed to be sine curves with one cycle, and a wave propagation algorithm is applied considering the randomness and irregularity of the materials and sewing parameters. The shape parameters are sufficient to represent the realistic shape of seam pucker. In order to construct neurofuzzy engines for a new evaluation method of seam pucker with the shape param eters, the engines are trained using many reference puckered seams for which the shape parameters are specified. The reference puckered seams are produced with a simulator when the shape parameters are specified.