The prediction of the surface weather variables like relative air humidity and wind speed and direction is a challenging task. The aim of this study is to evaluate the degree of importance of the surface weather variables used Numerical Weather Prediction (NWP). This evaluation was employed in the Consortium for Small scale Modeling (COSMO) model with grid box resolution of 2.8 km, in Muscat and Salalah International airports with specific and significant environmental and atmospheric situations. Based on the validation of 2 m height of relative air humidity and 10 m height of wind velocity, various adjustments were done to the model configuration. The weather surface parameters required to reduce the uncertainty of the COSMO model were determined in two stages.. The first stage consists of a 78-hours forecast with two different periods, 28-31 July 2019 and 8-11 February 2020. This stage showed that 9 out of 17, model parameters and dataset changes, have significant impact on the simulation. These parameters include, climatological data based on a better orographic dataset, soil type and topography, the reference temperature on sea-level, value for diffusion coefficient of heat and momentum, correction factor for horizontal length scale, scaling factor for the laminar boundary layer for heat, model time steps, and the subgrid-scale orography scheme. The modifications of the topography dataset improved the prediction of Salalah output but not that of Muscat and in particular the 2 m temperature. Using the information gained from the 78-hours forecast in stage 1, the one-month forecast was run in stage 2, which covers two periods, July 2019 and February 2020. Stage 2 runs were conducted once with inclusion of modification of topography and once without modification. It was found that computation without topography alteration provides better forecast that is comparable to the measured data.