The function of social media is now transforming to become source of information, even to support electronic word of mouth (e-WOM). Many companies, including ride-hailing service providers, can capture customers' opinions for the purpose of evaluating their products and services. Text mining can be useful to analyze great number of comments from ride-hailing customers in social media. Furthermore, by applying sentiment analysis, service providers can define the service categories which are good and still needing improvement. Customers' comments were taken from Twitter, and text classification method was used to classify the comments based on six predefined categories and their respective polarity. The accuracy of the classification model was 86% which was good to classify the text data. The output of this research is expected to give insight for ride-hailing service provider to understand customers' perspective about the services so that it will be easier to evaluate and improve their services based on the categories in this study.