Machine learning has emerged as one of the most influential branches of artificial intelligence and has experienced remarkable growth in scientific production during the last decade. Despite the increasing volume of research in Iran in the field of machine learning, there are limited scientometric studies on the knowledge structure and thematic evolution of Persian-language research in this domain. Therefore, this study aims to map the knowledge structure and thematic evolution of Persian machine learning research indexed in the Islamic World Science Citation Center (ISC) during the period 2013–2024. This research adopts a descriptive-analytical scientometric approach. Data were retrieved from the ISC database, resulting in 399 relevant articles after data cleaning and content validation. To improve the accuracy of the analyses, keywords and institutional affiliations were standardized manually. The data were analyzed using Biblioshiny and VOSviewer. Scientometric indicators such as scientific production trends, three-field plots (authors–journals–keywords), university collaboration networks, keyword co-occurrence analysis, thematic evolution maps, and strategic diagrams were employed to investigate the intellectual and thematic structure of the field. The results reveal a substantial increase in scientific production, from only one article in 2013 to 106 articles in 2023, indicating the rapid expansion of machine learning research in Iran. Among the 136 academic institutions that had articles in the field of machine learning, Islamic Azad University (101 articles), the University of Tehran (65 articles), and Tarbiat Modares University (23 articles) were identified as the most productive contributors. Furthermore, among the 1,873 identified keywords, "support vector machine," "random forest," and "artificial neural network" were the most frequent research topics, with frequencies of 60, 59, and 43, respectively. Thematic evolution analysis indicated a gradual transition from foundational machine learning concepts and traditional algorithms toward more specialized applications, including deep learning, modeling, landslide prediction, and geospatial analysis. Co-occurrence analysis of keywords led to the identification of four main thematic clusters in the field of machine learning: environmental and natural resource applications, methodological and algorithmic developments, interdisciplinary applications, and remote sensing-based studies. The strategic diagram identified “classification,” “prediction,” and “artificial intelligence” as fundamental themes with high centrality, whereas clusters related to “support vector machines,” “random forests,” and “remote sensing” were categorized as highly developed but specialized themes. The findings suggest that Persian-language machine learning research in Iran has entered a stage of relative growth characterized by a transition from algorithm-centered studies to application-oriented and interdisciplinary research. The field demonstrates a stable intellectual foundation while simultaneously expanding into new domains of application.