Sentiment analysis plays an important role in measuring public’s emotions and social media moods using affective tone evaluation of textual material. Sentiment analysis of text in Arabic is particularly challenging due to the diverse morphology of the language, right-to-left writing system, and dialectal variabilities. It is postings from Arabic social media regarding the 2023 Moroccan earthquake that are the subject matter of this research. We collected a large corpus of social media tweets, conducted preprocessing, and applied common machine learning models coupled with neural networks for sentiment analysis. Our results indicate that common models, particularly the Support Vector Classifier (SVC), are superior to deep learning techniques. These findings indicate the effectiveness of machine learning in Arabic sentiment analysis and contribute to the creation of Natural Language Processing (NLP) solutions for low-resource languages.