This paper investigated shear-related roughness classification based on Fuzzy comprehensive evaluation, and established a new strength model of natural rock joint. The joint profiles were extracted from 3D data to evaluate roughness, and a series of direct shear tests were carried out on three kinds of natural rock joints. The height statistical parameters Sm, Sq, Ss and Sk, and the textural statistical parameters Si, Sc, and Z2 were measured by Talysurf morphology instrument. Considering the existence of strong positive relations between Sm and Sq, and Z2 and Si, the four morphological parameters of Si, Sq, Sc, and Sk are considered as the main influencing factors related with rock joint shear behaviors. A new fracture roughness coefficient FRC is proposed based on fuzzy comprehensive evaluation (FCE), which considers that the influence of multi morphology parameters on the roughness of rock joint surface, and generally, the FRC is higher than the JRC. Compared with the previously published shear strength models, the proposed FRC-JCS shear strength model, which can more comprehensively reflect the influence of joint surface morphology parameters to shear strength, are in better agreement with the experimental data.