This study introduces the first hybrid annotation model for Arabic argumentation debate corpus. It aims to analyze Arabic argumentation structure in this competitive debate corpus using this hybrid model. The model combines Aristotle’s three appeals of logos, ethos and pathos with Toulmin’s model of argument structure analysis, in addition to some added labels inspired by the Arabic debate corpus. These added labels are self- repetition, team repetition and reciting, which were added to the hybrid model to reflect the uniqueness of the Arabic argumentation dataset. In addition, the hybrid model further subdivides Toulmin’s ‘backing’ into evidential backing and rational backing reflecting patterns emerging from the dataset annotation. The study presents preliminary findings, and significant patterns of the used labels in Arabic argumentation. The model addresses a gap and an under-representation of Non-English language argument annotation models by enhancing linguistic diversity in argument structure analysis, argument mining and Natural Language Processing (NLP). Thus, the model eliminates Artificial Intelligence (AI) bias by supporting a low-resource language such as Arabic. The multifaceted nature of the corpus on which the model was used, in terms of topics, gender representation, and geographic origins of debaters, serves as a robust resource for validating the model and for analyzing argumentative discourse in the Arabic speaking world. The implications of applying this hybrid model in Arabic argumentation research will be discussed.