In recent times, social media comments and tweets have proven to be a rich resource for the understanding of people’s opinions. The research done in this domain mainly focuses on sentiment analysis of tweets for simplistic multi-class classification. Conventional techniques are used to determine the polarity of sentiment namely positive, negative, and neutral. In this paper, we conduct a comprehensive comparative analysis of various models that are used to perform sentiment analysis, focusing primarily on tweets during the pandemic period of COVID-19 for multi-label emotion classification and emotion detection. Our work focuses on performing sentiment analysis on multi-label multi-class tweets taken from Twitter by using different models like LSTM, Naive Bayes, ANNs, and Transformers. The aim of this study is to give a comparative viewpoint of the various machine learning paradigms for multi-label classification for sentiment analysis. Different pre-processing techniques like GloVe, Word2Vec, CountTokenizer and Transformers as well as standalone Transformer models have been used and the results have been compared based on the model’s classification performance and precision as evaluation metrics. Our survey shows that pre-trained transformer based Large Language Models have performed the best with a maximum precision score of 71% during classification.