ABSTRACT Mine developments are the main access to extract tabular ore deposits in deep underground mines. Therefore, their stability is considered the principal priority during the mine production plan. The success of ore extraction mainly depends on the stability and serviceability of mine developments. Mine development instability is expensive and is a risk to personnel and equipment and in turn, it raises operational costs (e.g, repair costs, slashing, rehabilitation costs, costs of adding secondary support, miners wages and delay of production) (Ellefmo, and Eidsvik, 2009; Abdellah et al. 2014a; 2014b; 2014c). This paper aims to develop a hybrid approach in which deterministic numerical modelling is integrated with probabilistic methods to estimate the probability of unsatisfactory performance (e.g. rating and ranking) associated with the instability of mine developments with respect to mining sequences adopting Rosenblueth’s Point-Estimate Method (RPEM). A three-dimensional, elastoplastic, finite difference model (FLAC3D) is created (Itasca, 2009). The results are presented and categorized with respect to the probability of instability and the mining stage.