This study aimed to develop and validate a scale of agentic engagement in massive open online courses (MOOCs). Initially, 14 items were derived through a literature review, from which 7 items were drawn through a Delphi survey. Sample sets were used to develop and confirm the constructs of the new scale proposed in this study. To examine exploratory factor analysis, a survey of 163 learners using the K-MOOC system was conducted. The items were clustered into three factors: agentic support requests, agentic learning strategies, and agentic learning construction. A survey was administered to 243 respondents to validate the scale. The data gathered were analyzed through confirmatory factor analysis and a reliability test. The results showed that the agentic engagement scale, with 7 items and 3 factors, was reliable and valid. Based on the findings, instructional strategies were identified to enhance agentic engagement in self-paced MOOCs.