Several researchers have attempted to investigate the effect of joint roughness during loading by simplifying and making different assumptions. Some researchers, such as Barton and Labuscher, have presented visual models, and the aim has been to use and adapt the surface roughness of the joint rock to the presented models and estimate the surface roughness of the joint rock. Some researchers, such as Tse and Cruden, Meyers, and El-Sudani, have used statistical methods to express the surface roughness of the joint rock using quantitative parameters. Although visual methods are simple and fast, they are associated with human judgment errors and the inability to examine the preference of the roughness of one surface over another, and the estimation of the joint rock roughness is subject to errors. In quantitative methods, the use of a quantitative parameter cannot correctly express the characteristics of the surface roughness of the joint rock. In this paper, using field observations of the condition of rock joint surfaces and information collected from the surfaces available in the sources, the definition of possible geometric shapes of rock joints for 203 samples has been attempted. Possible geometric shapes were examined and the most commonly used shapes were defined in the form of 20 geometric models as representative geometric models SF, Rp, Rq, and Z2. They were evaluated and classified using quantitative parameters for determining roughness such as representative geometric models. In this classification, a more accurate estimate of the degree of joint roughness can be achieved by using the joint shape (similarity of the joint surface to representative geometric models) and considering the joint surface height parameter (K).