Background: Learning Arabic is often considered difficult by students because it involves various language elements that must be understood simultaneously. From the perspective of John Sweller’s Cognitive Load Theory, this condition is associated with the high demands of information processing in working memory, which can generate cognitive load. The study of cognitive load in learning Arabic is important for understanding students’ learning processes. Purpose: To describe students’ cognitive load during learning Arabic on the topic of al-mihnah in class X E at MAN Sumenep. Method: This study employed a descriptive quantitative, using all tenth-grade students of MAN Sumenep as the population, with 36 students from class X E selected as the research sample. The research instruments consisted of students’ worksheets (LKS) to measure intrinsic cognitive load and questionnaires to assess extraneous cognitive load and germane cognitive load. The questionnaire instruments were developed by adapting the multi-item indicators of Cognitive Load Theory proposed by Leppink and Klepsch and were constructed based on the factors and effects of cognitive load in the learning process. The instruments were translated and adjusted to the context of Arabic language learning and were subsequently tested for validity and reliability prior to data collection. The data were analyzed using descriptive statistical techniques to identify and categorize the levels of intrinsic, extraneous, and germane cognitive load experienced by students during Arabic learning. Results and Discussion: The study found that all three components of cognitive load were in the medium category. Intrinsic cognitive load had a mean score of 72.62, indicating that students were fairly capable of understanding the material and completing the worksheet tasks. Extraneous cognitive load had a mean score of 59.79, indicating that the instructional design sufficiently supported the learning process, although some unnecessary cognitive load remained. And germane cognitive load obtained a mean score of 71.13, indicating that students were fairly active in constructing and strengthening knowledge schemas. Conclusions and Implications: Students generally understand the material and constructed knowledge effectively, although the cognitive demands still require further optimization. These findings imply that Arabic language instruction should be designed in accordance with Cognitive Load Theory principles, with structured material presentation, appropriate media use, and clear instructions to reduce extraneous cognitive load and optimize students’ working memory capacity.