| As global demand for clean energy technologies continues to rise, the exploration of critical minerals has become a top priority. These include elements such as lithium, nickel, cobalt, uranium, and rare earth elements, which are key components in electric vehicles, batteries, wind turbines, and many advanced technologies. This thesis examines the significance of computational methods in mineral exploration through three main contributions. The first is a thorough review of machine learning (ML) techniques utilized in this domain. It covers supervised and unsupervised learning, deep neural networks, and Generative Artificial Intelligence (GenAI), applied to various geoscientific datasets. The second contribution explores lithology classification using ensemble learning techniques such as XGBoost, CatBoost, and Random Forest. This is applied to the FORCE 2020 dataset, encompassing 118 wells from the Norwegian North Sea. The final component focuses on lithostratigraphic well log correlation using a modified version of the Fast Dynamic Time Warping (FastDTW) algorithm. While ML shows promise, challenges like data scarcity limit its application. Hence, the modified FastDTW algorithm is applied to a carbonate reservoir dataset from southwest Iran, achieving significant performance improvement over a Principal Component Analysis (PCA)-based method employed in a recent research. Overall, this thesis underscores the importance of both ML and traditional methodologies in advancing critical mineral discovery. Keywords: Critical Mineral Exploration, Machine Learning, Lithology Classification, Lithostratigraphic Correlation, Ensemble learning, Dynamic Time Warping |