A chatbot is a computer program that can understand the human language and respond like humans. The more it acts like a human, the more helpful the chatbot is. There are several attempts to make the chatbot intelligent or human-like. In college, the routine task is to answer the questions that have been asked repeatedly or found on the official website. This task is time-consuming and a wastes task. The chatbot is one of the solutions to this problem. A good chatbot can naturally answer such questions quickly and tirelessly. However, producing an intelligent or human-like chatbot is a challenge. The chatbot should be able to answer the fundamental questions in the domain and the advanced opinion questions. In this research, we proposed the chatbot system using the machine learning technique. Therefore, the chatbot can learn from the user to improve its performance. We evaluate our chatbot by testing with real-world conversation. The result shows that our chatbot can respond to the fundamental questions with a higher accuracy rate than the advanced questions.