Natei ErmiasEyale BayableYedilfana SetargeKefelegn Shewangzaw2025-08-312025-08-312024-02https://etd.aau.edu.et/handle/123456789/7195Because of the natural fluctuation in wind energy availability between the dry and wet seasons, where hydropower reservoirs experience varying water levels, wind power can serve as a vital supplement to hydroelectric power within Ethiopia's energy framework. Because wind power increases system reliability even during dry seasons, it becomes an essential component of the grid energy mix. This study has focused on predicting wind energy potential in the mountainous Wello highlands region upper basin of Abay and Tekeze of Ethiopia using Artificial Intelligence (AI) and empirical models. Wind speed data from four locations (Wegertena, Lalibela, Amdewak, and Ambamalyam) and eight years data (2013-2020) was used to develop AI and Empirical models for the prediction of wind energy. Long short-term memory (LSTM) neural networks provided the highest accuracy in modeling wind speed, capturing complex seasonal patterns and explaining more than 97% of the variance. On the side of wind power density, all four models at all study sites perform very accurately, scoring above 98.8%, but the gradient boost is the advanced model from the other models at 99.9% only wind power density simulation. The evaluation metrics consistently show that the AI model outperforms the empirical methods on all tasks. LSTM architectures can incorporate long-term weather context, resulting in significantly improved predictions compared to snapshot models that lack native sequence understanding. Accurate wind energy forecasting supports the expansion of renewable infrastructure to sustainably meet Ethiopia's rapidly increasing electricity demand. This study also demonstrates the effectiveness of modern AI in the complex spatiotemporal modeling of wind resources. This provides evidence that deep learning approaches such as LSTM are uniquely positioned to capture diverse wind dynamics based on real-time time series dependency modeling. Accurate wind energy forecasting will be critical as Ethiopia pursues an ambitious sustainable development and decarbonization path over the coming decades. Integrating AI-powered platforms into renewable energy planning and operations will facilitate optimized wind farm site selection and grid integration toward national green energy goals.en-USAIANNXGboostGradient BoostLSTMWello HighlandsWind Energy Forecastingand Weibull ParametersThesis