 The study aimed to develop machine learning and transfer learning models to classify tweets related to COVID-19 vaccines and explore temporal trends and topics across a large dataset of tweets from November 2020 to January 2021. The results showed that the transfer learning models outperformed the machine learning models in detecting opinions, attitudes, and behavioral intentions towards COVID-19 vaccines. Additionally, the prevalence of positive behavioral intentions increased significantly in December 2020, and the study identified 10 main topics and relevant terms for each category of tweets related to COVID-19 vaccines. The findings can be used to tailor educational programs and other interventions to promote public acceptance of COVID-19 vaccines. This article was authored by Siru Lu, Geely Lee, and Jelin Lu.