Various roof support design methodologies have been used in Australian coal mines, which include analytical, numerical and empirical models. These models are mainly based on the deterministic approach in which a single factor of safety is calculated for the roof support design. The main limitation of this design methodology is that it fails to account for the inherent variations existing in rock mass properties and other roof reinforcement elements. To overcome this issue, an improved design methodology based on stochastic approaches has been developed in which both the design inputs parameters and the outcomes (i.e., factor of safety) are expressed as probability distribution functions. This paper focuses on the application of stochastic modelling technique to evaluate the underground roof support strategies currently used in an underground coal mine located in the Bowen Basin. The starting point of the analysis is the existing analytical roof support models that identified the relevant design inputs in consideration. Based on the best fit probability distributions of input parameters determined by goodness of fit tests, a risk based design is conducted to quantitatively evaluate the risk of roof fall fatality under specific roof support system by using the probability of failure from Monte Carlo simulation and the associated underground personnel exposure.