There is not any unique solution when stability of a slope in a fractured rock mass is evaluated due to some degree of uncertainties. Uncertainty occurs due to variation in mechanical properties of intact rock and discontinuities, analysis tools and also geometrical properties of fractures. Therefore, probabilistic analysis is inevitable. In the literature in slope stability problems, they have considered continuum/equivalent continuum models based on the calculated mechanical properties of rock mass using empirical methods or just considered regular joint set system as the inpu t geometrical parameters of discrete models. Stochastic DFN models present a more realistic representation of fracture network pattern. The DFN models are based on stochastic representations of fracture system, using the probabilistic density function of fracture parameters using Monte Carlo model, formulated according to field mapping results. The force acting on blocks also is estimated using numerical methods such as Distinct Element Method (DEM) approach. A methodology development for a numerical probabilistic analysis of slope stability using DFN-DEM models is the main objective of this research work. In the right abutment of Karun 4 dam in Iran and in the downstream of the dam body, five joint sets and one major joint have been detected. According to the geometrical properties of fractures in the Karun river valley, there is some potential unstable condition for this abutment. A number of stochastic DFN models are generated and a series of numerical DEM modeling are performed on generated models to calculate potential block failure and measuring minimum required support patterns for dry and assumed water table static loading and also dynamic analysis. The dynamic input data is based on two expected Design Basis Earthquake (DBE) and Maximum Credible Earthquake (MCE) conditions. Present slope condition at right abutment in Karun 4 hydro power plant is well much with the results of slope stability analysis with dry condition. Results of dry and assumed water table conditions show that the distribution of required bolt lengths follows a lognormal distribution. In dry condition, more than 80 percent of models need only 1614.99 m of bolts which is a bolt pattern with 2 m spacing and 12 m of length. Stronger bolt pattern with 2653.49 m of bolts which is a bolt pattern with 15 m length and 1.5 m spacing for assumed water table condition is needed for 80 percent of models. Dynamic analysis results indicate that MCE condition is likely to trigger sliding rock mass on major joint of slope. It shows also the failure of blocks on major joints is occurred locally in DBE condition and consequently it leads to the slope to be unstable in the future.