The case study on strong temperature inversion event in Armenia has been presented in this paper using observed temperatures and simulations from both global (ERA-Interim) and regional high-resolution (WRF) models. According to the observations the December of 2013 was ranked as the second coldest since 1961. Monthly mean temperature anomaly for entire Armenia consisted of -4.4 °C in December, while that for inversion affected basins such as Ararat valley was -9-7 °C. The persistent temperature inversion in Ararat valley lasted more than two weeks resulted in unprecedented cold wave events. The ability of the global ERA-Interim and regional WRF models to simulate this dramatic temperature event in Armenia has been examined. The results show that high resolution WRF model has a clear advantage over ERA-Interim model in representation of spatial temperature pattern in Armenia. The small-scale variations of temperature and surface inversion basins are simulated by WRF model while ERA-Interim model provides very coarse results underestimating the spatial variability and influence of topography. However, it should be noted that there are significant uncertainties and errors in WRF temperature forecasts. Significant RMSE values and negative correlation coefficients obtained for the area covering inversion basin indicate that WRF model fails to capture the temporal variability of temperature during the inversion event. It is worth noting the significant positive bias for daytime WRF temperatures during the strongest phase of the inversion.