Event Extraction has emerged as a research hotspot in Natural Language Processing, with its applications spanning diverse domains such as newswire, biomedical sciences, history, humanities, and cyber security. Addressing this complex task, numerous methods, datasets, and evaluation measures have been proposed in the literature, spanning multiple languages and, less frequently, Arabic. The distinctive challenges posed by the Arabic language, including its morphology, syntax, ambiguity, and scarcity of annotated data compared to languages like English, underscore the necessity for an exhaustive and up-to-date survey. In contrast to several survey articles that primarily focused on defining the task, evaluating methods, benchmarking datasets, and exploring deep learning-based solutions, our approach not only summarizes task definitions, data sources, and performance evaluations but also offers a taxonomy of methodologies for event extraction. Furthermore, we provide a comprehensive analysis of the most representative methods within each approach, delving into their fundamentals, strengths, weaknesses, and principal applications in the Arabic language. Lastly, we present our vision for future innovative research directions, with a particular emphasis on approaches leveraging Pre-Trained Language Models.