基于分形几何的码尺法分维数与岩石节理粗糙度系数的物理意义剖析，认为Ｄ－ＪＲＣ之间不存在必然的相关性．分析标准轮廓曲线的分维数，发现其分维数差级微小，难以实行粗糙度系数分级．根据实测资料阐述了岩石节理表面轮廓曲线的“自相似”是统计意义而不是绝对的，它要求ＪＲＣ分形估测应统计求取，而过繁的分维数测量步骤削弱了ＪＲＣ的分形统计估测的可行性．建立在实测资料统计分析基础上的ＪＲＣ尺寸效应分形模型ＪＲＣｎ＝ＪＲＣ０（Ｌｎ／Ｌ０）－Ｄ客观而真实地刻画了粗糙度系数随取样长度增大而降低的规律，其中，ＪＲＣ尺寸效应分维数（Ｄ）具明确的物理意义，它描述了ＪＲＣ随结构面规模增大而降低的衰减速率．最后，运用ＪＲＣ尺寸效应分维数（Ｄ）探讨了岩石节理粗糙度系数尺寸效应的各向异性规律． Based on fractal geometry code dimension fractal dimension and rock joint roughness coefficient of the physical meaning analysis, that there is no inevitable correlation between the D-JRC. Analyzing the fractal dimension of the standard profile curve, we find that the fractal dimension of the standard curve is very small and it is difficult to apply the roughness factor classification. According to the measured data, it is stated that the “self-similarity” of the rock profile curve is statistically significant rather than absolute. It requires that the JRC fractal estimation should be calculated statistically, while the excessive fractal dimension measurement steps weaken the JRC fractal statistical estimation Test the feasibility. JRCn = JRC0 (Ln / L0) -D based on the statistical analysis of measured data objectively and truly depicts the rule that the roughness coefficient decreases with the increase of sample length. JRC size effect fractal dimension The number (D) has a clear physical meaning and it describes the decay rate at which JRC decreases as the scale of the structural plane increases. Finally, the anisotropy law of size effect of rock joint roughness coefficient is discussed by using JRC size effect fractal dimension (D).