Jointed rockmass is involved in many civil engineerings. Correct assessment of the shear strength of rock joint surfaces is essential for such projects design. However

there are various factors to affect shear strength of rock joints. Among them

the surface shape of the structure plane is vitally important for its mechanical property because the surface shape affects the real area of contact and frictional status definitively. In this study

the rough joint surface is considered to be made up of a series of microcosmic rectangular-shaped asperities with different heights. There might be two failure modes of a rectangular asperity: a dilative failure mode and a non-dilative failure mode

although the rough joint surfaces show macroscopic shear-dilatancy failure in most cases. Based on them

the stochastic strength model of rough joint surfaces is established by applying the probability density function to describe the asperity height distribution. The shear strength is theoretically derived according to the proposed model. Direct shear tests on artificial rough joint surfaces are performed to verify the proposed stochastic model and its application method. The theoretical computations of stochastic model provide a good agreement with the obtained peak shear strength and residual shear strength from shear tests

which may serve as an effective approach for quantitatively estimating the peak shear strength and residual shear strength of rock joint surfaces.