Taxation is the main state data, to make it easier for people to pay taxes, an integrated application called coretax is used. However, there are many differences because it has many problems in its use. Therefore, the study was conducted to find out public sentiment towards the coretax application using a naïve Bayes algorithm. The methods used range from data collection, data cleaning, data pre-processing, classification with textblob to classification and evaluation with naïve Bayes algorithms. Of the total 2858 total data used, the results were 782 positive sentiment data, 479 negative sentiments and 1597 neutral sentiments. The results showed that the accuracy of the model could reach 81% with an f1-score value of 80%, as well as a precision and recall value of 81%. This shows that the naïve bayes algorithm is quite good at classifying public sentiment towards the coretax application.