Fake news stories can polarize society, particularly during political events. They undermine confidence in the media in general. Nowadays, it has never been easier to disseminate “fake news” especially through social networks. Given the high stakes involved, determining truth in social media has recently become an emerging research that is attracting tremendous attention. Current NLP algorithms are still missing the interpretation of the news. Therefore, using approaches based on news content challenges the detection of fake news’ characteristics; we need auxiliary information to help make a determination. ‘Social-context-based’ and ‘propagation-based’ approaches can be an alternative or complementary strategy to content-based approaches. This paper surveys the work on Arabic fake news detection methods and techniques. Particularly, we identify the details related to the fundamental theories across various disciplines by situating their usage within current opinion mining techniques in social media to encourage interdisciplinary research on Arabic fake news. An exhaustive review of Arabic fake news detection techniques on social media is presented, including the task definition, the different approaches and the various deep learning approaches applied. This survey can facilitate the collaboration of efforts among experts from different disciplines such as in computer sciences, social sciences and political science to research fake news. In addition, this study explores the emerging trends and the different applications of Arabic fake news detection on social media. The study concludes by providing discussion of the gaps in the current existing research and highlighting the possible future directions for stance detection on social media.