Due to the intermittent nature of wind, wind resource assessment and electrical power forecast require an accurate wind data. Numerical Weather Prediction (NWP) models are used to generate the required data. Different parameters such as the initial state of the atmosphere and the boundary conditions play major roles in the quality of the generated wind data. Therefore, the provision of an accurate wind data presents a major challenge to wind energy market. This thesis aims to propose and develop approaches for better use of NWP models in wind energy applications. All proposed approaches were validated over the Sultanate of Oman. In resource assessment aspects, NWP nested ensemble approach was proposed to develop wind parameters database at high resolution (2.8km) and at different heights above the ground. Parameters include wind power density, wind speed frequency distribution, turbulence intensity and peak hour matching. Model validation results showed that the Bi-linearly interpolated ensemble mean performed better on average than each individual ensemble member. Using the generated wind parameters database, multi-criteria decision making technique was used for wind farm land suitability analyses. This technique was based on fuzzy logic using the aggregation operator extension (AHP-OWA combination) for wind resource assessment. This approach aims to select the best site for implementing wind farm projects. Based on the proposed approach, wind farm suitability maps for Oman were developed for different heights above the ground. Oman land suitability maps showed that the mostly suitable sites for wind farm installation are located on the area between Thumrait and Qayroon Hirity at Dhofar region and also located on the south eastern coast of Wusta region near Ras Madrakah and Ras Qarwaw. After selecting the best site for the wind farm, it is necessary to optimize the layout of the wind farm to maximize the energy production. Therefore, a geometrical approach for wind farm layout optimization was developed. This approach was used to determine the optimum orientation angle that will minimize the wake effect and hence maximize the capacity factor of the wind farm. It was demonstrated that the optimum angle used for orienting the vertical axis of aligned columns of the turbines relatively to the predominant-wind median axis should be around 18°. Selecting the best site doesn't guarantee continuous power generation from the wind farm. To assure an acceptable security of supply level and economical dispatch of the generated power, it is necessary to estimate in advance the expected power generation from the farm. It is also important to quantify the uncertainty of the generated power forecast. In this aspect, a spatial-temporal neighborhood approach for generating probabilistic wind and power prediction was proposed. This approach was used to convert the deterministic NWP model forecast to probabilistic forecast. In addition, this approach was used to quantify the uncertainty on the wind speed forecast. Validation results showed that spatial-temporal neighborhood approach was able to reduce the root mean square error by more than 50% compared to the deterministic approach over the validation sites