Fake news is untrue information presented as news. It often has the aim of damaging the reputation of a person or entity or making money through advertising revenue. It is arguably one of the most serious challenges facing the news industry today. Especially in the context of the neo-coronavirus epidemic, the truthfulness of the news is important. The goal of the Fake News Identifier is to explore how to use machine learning to process fake news. We're working on fake news classification is modelled as a binary classification problem. Find in the words and contexts that related words appear in the text and how to classify them as true news or fake news. Assessing the veracity of a news story is a complex and cumbersome task, the initial goal is to uncover the features embedded in a collection of fake news stories. Then model the data to find out the key features of the classification, and use the machine learning and deep learning model to make classification predictions and determine if it is fake news. After comparing 6 machine learning models and exploring the neural network changes (standard LSTM, two-way LSTM), it is found that the two-way LSTM can achieve the highest accuracy with the smallest overfitting, but the advantages are not obvious compared to other models.