I found some articles on the subject. Maybe these can help you.

Article Fake News Detection in Arabic Tweets during the COVID-19 Pandemic

Article Fake Information Analysis and Detection on Pandemic in Twitter

Article Using Social Media to Detect Fake News Information Related t...

https://europepmc.org/article/med/34611504

I suggest you look at kaggle, e.g., https://www.kaggle.com/datasets/clmentbisaillon/fake-and-real-news-dataset and https://www.kaggle.com/c/fakenewskdd2020 etc.

Check out the latest Stanceosaurus dataset which provides annotated tweets for 3 languages (Arabic, English, and Hindi):

Preprint Stanceosaurus: Classifying Stance Towards Multilingual Misinformation

Thanks all for your reply and your attention which is admirable for me...

I will follow your suggestions to find any solution, but the biggest problem in our research is the lack of data (annotated) in persian language. I will continue my research to find that but do you suggest unsupervised methods in the other hand for Fake News Detection if there is no annotated data? (that's not promising for me for best solution)

Mehmet Karahan

Muhammad Ali

Tarek Naous

You can try to label a few data samples in Persian and use them for few-shot in-context learning using a Persian pre-trained language model (if available) or a multilingual language model.

Tarek Naous

Thanks again dear Tarek, Yes this solution is good because there is some pretrained models based on BERT for our language but the challenge is labeling because of understanding Fake or Real news on twitter and for this task we need experienced experts who can label a tweet containing news as Fake or not. but we continue researching about.