Abstract: A systematic scheme on fabric point-sampled model was introduced to objectively assess the fabric smoothness grade.The high precision laser scanner was adopted to obtain the fabric geometry.A point cloud composed of scatter three-dimensional coordinate points was the discrete representation of the fabric surface.Feature points were extracted from the point cloud and neighborhoods at vertices were constructed.The calculations of Gaussian curvature and mean curvature were performed later using surface discrete differential geometry.The statistic results show that Gaussian curvature,mean curvature and their combination estimate the local surface crooked properties and bending type.The digital analysis on fabric point-sampled model supplies believable and reliable data for the objective assessment of fabric smoothness grade.