Beginning in late 2019, the COVID-19 pandemic had a significant effect on people and communities all across the world, influencing not only public health but also a number of societal facets. Sentiment analysis, a branch of natural language processing, was essential in helping to comprehend how people's feelings and perspectives changed throughout this world crisis. Using machine-learning techniques, this research provides a thorough summary of sentiment analysis during the COVID-19 pandemic. Large-scale textual data from social media, news stories, and public forums were gathered and analyzed for this study. To identify the sentiment in these texts, a variety of supervised and unsupervised machine-learning techniques were used. To identify emotions like fear, anxiety, hope, and disinformation, sentiment analysis models were developed and put to the test. The study also looked at the difficulties brought up by the talks surrounding COVID-19 being so dynamic and the requirement for ongoing model modification. The results show that sentiment has changed over time, mostly in line with how the virus has spread. There was a noticeable spike in anxiety and panic early in the pandemic, but as society adjusted to the new normal, these feelings progressively gave way to hope and fortitude. The study also highlights the significance of sentiment analysis in tracking and preventing the spread of misleading information and identifies disinformation as a major threat. The study's findings show how effective machine learning and deep learning techniques are in monitoring and interpreting public opinion in times of international emergency. Researchers, legislators, and public health authorities can use this information to help them create successful communication plans and make well-informed decisions. Sentiment analysis is still a useful instrument for assessing public opinion, responding to issues, and building resilience in the face of hardship as the globe struggles with the COVID-19 pandemic.