Understanding the emotions expressed in text has become essential for businesses in the age of digital communication in order to assess client mood and adjust their offerings. But it is still quite difficult to decipher emotions from text accurately. In order to tackle this issue, the study dataset, obtained from Kaggle, consists of 16,000 Twitter records, each linked to one of six distinct moods. The effectiveness of classification is examined for the following models: Decision Stump, Hoeffding Tree, J48, Naïve Bayes, Naïve Bayes Multinominal, and Naïve Bayes Updateable. With this, we present an overview of the architecture of our suggested system, which combines natural language processing and machine learning algorithms, through an extensive analysis of previous research and approaches.