Machine Learning is essential for many applications, from personalized recommendations in entertainment apps to medical diagnostics and robotics. It uses algorithms to synthesize relationships in data. Given the data’s increasing volume and dimensionality, a crucial step is feature selection, which aims to reduce this dimensionality by retaining only the most relevant information without losing efficiency. Various techniques for feature selection have been developed, and new ones are regularly proposed. This systematic literature review evaluates the use of Rough Set Theory in feature selection techniques utilized between 2018 and 2023. Numerous studies present techniques developed using rough set theory derivatives, from traditional reductions to combinations with hybrid, dynamic, parallel methods, three-way decision techniques, neighborhood methods, etc. Previous reviews have provided comprehensive theoretical frameworks but often lacked detailed discussions of review results. This article proposes an extensive and detailed systematic literature review on applying Rough Set Theory in feature selection to determine the state-of-the-art models that show the best results in their respective applications.