Taking the rock nature joint surface as the research object, the morphology of joint surface was measured and 3D shape of the joint surface was established. On the basis of this, five parameters were calculated and the correlation between parameters was studied taking the root mean square value（RMS） as benchmark. Selecting three morphology parameters, the method of fuzzy mathematics was used to quantify the roughness of joint surface, and the roughness was classified. Meanwhile, the correlation between the morphology parameter and roughness index was studied. The results show that the average height of the central line is proportional to its RMS. The relation among the ratio of kurtosis coefficient, skewness coefficient, micro convex angle and RMS of mean height of centerline is power functions. The results of quantitative classification of roughness show that the quantitative classification index and friction coefficient and friction angle have obvious positive proportional linear relationship. It can be used for the estimation of joint surface friction. The relation among the ratio of friction coefficient, friction angle, the quantitative classification index and RMS of mean height of centerline is also power functions. Through the quantitative function relationship between the parameters, the correlation coefficient can be obtained, and then the estimation of surface roughness and friction force of the joints can be realized.