Coffee is an Indonesian plantation product that has high competitiveness in the international market. The demand for coffee is influenced by people's perceptions of coffee in Indonesia. People easily provide opinions and opinions through social media. The most popular social media application in all circles is Twitter. The types of coffee that are trending topics on Twitter today are civet coffee, black coffee, sick coffee and bitter coffee. The purpose of this study is to provide an overview of the positive and negative public sentiments towards coffee in Indonesia. In this research, sentiment analysis using Naïve Bayes Multinomial has been applied to determine the perception of good or bad in society. The results showed good accuracy such as accuracy, precision, recall of 94%, 99%, 88%. The results of this study prove that the proposed model can be used to analyze sentiments on Twitter texts very well.