The social networking site and user sites contain a large amount of information with users' feelings and opinions in various fields. For example, pharmaceutical companies provide users with text reviews and drug numeric ratings. Anyhow, these text-oriented reviews may not always be consistent with numerical values. In this project, we use different sentiment analysis models to differentiate drug user rating emotions and compare their accuracy. Various machine learning models including Logistic Regression, XG boost, and Naïve Bayes classifiers are being implemented by pushing it with drug reviews as inputs. In this, the XG boost model performed much better than other models with a total accuracy of 79.7%. This study has shown that these classification models can be used to separate drug reviews and identify the overall polarity or sentiment of the consumers. Since the main focus of this study was to separate the text reviews, the data were segregated according to each polarity. The division of binary classification was chosen instead of multiple classes as the purpose of this project was to identify the best and worst points/polarity. With the help of this polarity, it helps the nutritionist and the doctor to get updated with the in-market polarity of a particular drug and its public response.