The quick advancement of technology in internet communication and social media platforms eased several problems during the COVID-19 outbreak. It was, however, used to spread untruths and misinformation regarding the illness and the immunization. In this study, it is examined whether machine-learning algorithms (Naive Bayesian, Random Forest, Logistic Regression, Decision Tree, and Support Vector Machine, as well as Gradient Boost, Bagging, AdaBoost, Stochastic Gradient Descent, and Multi-layer Perceptron) can automatically classify and point out fake news text about the COVID-19 pandemic posted on social media platforms. The “COVID19-FNIR DATASET” was used to train, test, and fine-tune machine learning models in order to predict the sentiment class of each fake news item on COVID-19. The results were assessed using a variety of evaluation metrics (confusion matrix, classification rate, true positives rate, etc.). The findings collected demonstrate an extremely high level of accuracy when compared to other models.Â