Purpose This paper examines the evolution of massive open online courses (MOOCs) literature over the past 15 years and identifies its significant developments.Design/methodology/approach Utilizing main path analysis (MPA) on a dataset of 1,613 articles from the Web of Science (WoS) databases, the authors construct the main pathway in MOOC literature through a citation analysis. Pajek software is used to visualize the 34 influential articles identified in the field.Findings Three phases emerge in MOOC research: connectivism as a learning theory, facilitating education reform and breaking barriers to MOOCs adoption. Multiple-Global MPA highlights sub-themes including self-regulated learning (SRL), motivation, engagement, dropouts, student performance and the impact of COVID-19.Research limitations/implications First, data limitations from the WoS core collection might not cover all research, but using reputable sources enhances data validity. Second, despite careful algorithm selection to enhance accuracy, there remains a limitation inherent in the nature of citations. Such biased citations may result in findings that do not fully align with scholars' perspectives.Practical implications The authors' findings contribute to the understanding of MOOCs literature development, enabling educators and researchers to grasp key trends and focus areas in the field. It can inform the design and implementation of MOOCs for more effective educational outcomes.Originality/value This study presents novel methodologies and important findings for advancing research and practice in MOOCs.