The development of vaccines and the widespread protection from vaccination have become the priority to end the epidemic. It is of great significance to study the public’s concerns and sentimental tendencies on the COVID-19 vaccine and then take effective measures to improve the vaccination rate. Nowadays, the processing of a prodigious amount of information is dependent on artificial intelligence. As a fantastic intersection of human and machine, an important research direction of artificial intelligence is NPL(Natural Language Processing). Through the analysis of social media data, it is conductive to make better business decisions, product optimization and human activity prediction. In this one-semester project, I analyzed the tweets about vaccines with deep learning and visualized twitter user’s sentiment. The project is divided into two modules: sentiment analysis and tweets data visualization. Specifically, in the part of sentiment analysis, I completed a three classification task (positive, neutral, and negative sentiment) with the language model ‘Bert’ (Bidirectional Encoder Representations from Transformers). The classification result is visualized by a word cloud and a sentiment map.