This study presents an optimisation model to achieve an optimal voltage sag monitoring programme. To cover the main uncertainty associated with fault impedance, a novel approach termed as probabilistic monitor reach area (PMRA) is developed. Moreover, PMRA is employed to consider reliability of the monitoring equipment impact. Additionally, several indices based on PMRA are defined to distinguish the optimal programme when outage of generators, loading level, line outage and feeder reconfiguration may result in several programmes. Besides, the other advantage of the proposed method is aptitude of modelling large load switching as an influential source of voltage sags. The model determines the number and the location of monitors in such a way that a pre-defined probability of observability is satisfied. Capability of the proposed methodology is checked for several test systems and illustrated here on the 24-bus reliability test system and on the 34-node test feeder. Numerical results and their comparison with the commonly used MRA method reveal that the proposed optimal voltage sag monitoring programme is robust against uncertainties including fault impedance, load variations and outage of the generators/lines.