Within the age of social media, stages like Twitter serve as a wealthy source of open conclusion, making estimation investigation an fundamental apparatus for understanding collective states of mind. This study employs both deep learning and machine learning models, such as Logistic Regression, CNN, and LSTM, for sentiment classification on Twitter data. After wide preprocessing and highlight building, we compare these models utilizing key execution estimations such as accuracy and F1-score. Our comes about show that CNN accomplished the most elevated precision at 98.64 beating other strategies. This examination highlights the prevalence of significant learning for large-scale supposition examination, underscoring its potential in numerous applications, from exchange bits of information to open conclusion watching.