AISnet For Students is an academic information system built by the Garut Institute of Technology to make it easier for students to carry out various campus academic administration activities online. This research aims to conduct sentiment analysis of online academic services at the Garut Institute of Technology by involving students as research subjects. This sentiment analysis will be carried out using the Naive Bayes Algorithm to explore student views and opinions regarding these academic services. This research was conducted with the aim of identifying potential problems that may occur in online academic services at the Garut Institute of Technology. Apart from that, this research also aims to provide recommendations that can help in improving the quality of these services. Research shows that students have positive sentiments towards academic services on campus. However, there are several problems that need to be overcome, such as technical problems and lack of features in the system. The solution to overcome this problem is to develop a user-friendly system, improve network quality, improve system features, conduct training or socialize the use of the system to students, and apply the latest technology and innovation in online student academic system services. The results of this research have the potential to provide benefits to educational institutions by helping to improve online academic services better. The results are expected to increase satisfaction and quality of services provided to students. Apart from that, this research can also be a reference or reference for further research related to sentiment analysis in the academic field or other fields. Where the Naive Bayes algorithm is used to analyze student sentiment towards academic services on the Garut Institute of Technology campus. The final results show that negative sentiment is greater than positive sentiment. Where negative sentiment is 54.75% and positive sentiment is 45.24%, this is because in the AISNet application most users provide reviews for the updates which are not real time. The following is the final result with an accuracy of 80.06%, a resolution of 83, 11 and recall 75.21.