Sentiment analysis is a technique used to analyze the subjectivity of opinions expressed in the text. In this research, we evaluate sentiment classification methods for analyzing public opinion about fuel inflation on Twitter, including Naive Bayes, Support Vector Machine (SVM), and Random Forest. Our results show that the SVM and Random Forest methods produced the highest accuracy rates of 78%, while Naive Bayes achieved an accuracy rate of 70%. Based on these findings, the SVM and Random Forest methods are good choices for sentiment analysis on public opinion about fuel price increases on Twitter.