<p>Microplastic pollution is a major global sustainability challenge, affecting ecosystems, human health, and progress toward the United Nations Sustainable Development Goals (SDGs). Although scientific output has grown rapidly, limited research has systematically examined the evolution, structure, and knowledge gaps of this field. This study addresses that need through a three-phase approach: (1) a bibliometric analysis of 12,544 publications from 1974 to 2025; (2) the application of machine learning techniques, including natural language processing, logistic regression, Random Forest, and XGBoost, to identify influential contributions; and (3) a systematic review of key studies published between 2021 and 2025, developed from the results obtained through the logistic regression model and conducted following the PRISMA protocol. Data were retrieved from Scopus and analyzed in R using the Bibliometrix package for text mining and data cleaning, complemented by the Tree of Science method to map citation-based knowledge structures. Results indicate that China, the United States, and India lead in publication output and methodological diversity. Despite this growth, important gaps persist, particularly the lack of standardized detection methods and insufficient research on interactions between microplastics and other pollutants. The findings underscore the need for integrative conceptual frameworks to reduce theoretical fragmentation and strengthen evidence-based policymaking. Overall, the field is shifting from impact assessment toward solution-oriented research, highlighting the importance of aligning technological innovation in degradation and resource recovery with environmental risk management and circular economy strategies.</p>