This study integrates unsupervised learning algorithms and fuzzy inference systems to improve decision-making in the oil and gas industry. The first part examines the application of clustering algorithms, such as K-means, Fuzzy C-means, and Hierarchical clustering, to well classification. Effective well classification is crucial for optimizing reservoir management and operational efficiency. Each algorithm’s performance is evaluated based on reservoir characteristics, well settings, and operational variables. K-means offers simplicity, hierarchical clustering provides structural insights, and fuzzy C-means facilitates soft clustering. These models are shown to enhance operational workflows and decision-making in the energy sector. The second part addresses the challenge of selecting candidate wells for extended shut-ins using fuzzy logic. Shut-ins, driven by economic or environmental factors, require a systematic approach for decision-making. A fuzzy expert system was developed using Python to select optimal wells for shut-in based on net present value (NPV). The system first clusters wells by performance and then recommends shut-in scenarios based on economic conditions. Its effectiveness is validated through simulations of a mature sandstone reservoir, demonstrating fuzzy logic’s utility in real-time decision-making under uncertainty. This integrated approach demonstrates the complementary roles of unsupervised learning and fuzzy inference in addressing complex operational challenges in the oil and gas industry. Future work may explore further refinements and broader applications to diverse reservoir types.