This research is motivated by the continued dominance of intralingual errors in the academic writing of BSA students despite their intensive Arabic language learning, particularly in the morphological and syntactic aspects that influence the quality of scientific argumentation. This study aims to identify, classify, and analyse the types of intralingual errors in the final assignment of the Kitabah course and explain their implications for the mastery of Arabic language structure as a basis for developing pedagogical interventions and curriculum evaluation in higher education. Using an interpretive qualitative paradigm with an instrumental case study design, data were obtained through document analysis, questionnaires, and interviews validated through triangulation, member checking, peer debriefing, and audit trails. The analysis followed the classic stages of error analysis. From 41 files, 138 errors were found, which were entirely intralingual and distributed across seven linguistic aspects, with error patterns influenced by internal learner mechanisms such as overgeneralisation, simplification, imperfect rule application, and rule overlapping. These errors disrupt the clarity of meaning, consistency of structure, and the accuracy of scientific argumentation. The study concludes that the source of the errors is not L1 interference, but instability of rule internalisation that requires pedagogical intervention based on error analysis. The study's limitations include its coverage of a single study program, document-based data, potential recall bias, and the lack of quantitative measurements. Nevertheless, this research is novel because it comprehensively maps intralingual mechanisms and offers an integrative analytical framework that can support improving the quality of Arabic language learning in accordance with SDG 4 and strengthen error-based teaching approaches in higher education.