With the exponential growth of user reviews with the development of the Internet, the restaurant field has gradually developed online under the rapid development of the Internet, and the review analysis system for users has become quite valuable for research. Although it is developing rapidly, there are some problems, such as the fact that most information on the Internet often contains a lot of noise, such as typos and emojis, which can interfere with the performance of the comment analysis system. The language in a review is often rich in emotion and complex semantics, and how to accurately understand and extract the key information is a technical problem. Based on the above problems, a restaurant review analysis system was proposed, which firstly extracted keywords and generated a word cloud map of key review words by collecting review data on review websites. Then, the sentiment analysis is realized based on SnowNLP in natural language processing. Finally, PyQt was used to deploy the front-end interface, providing users with a simple and convenient restaurant review analysis tool. This system solves the information deviation between the restaurant owner and the user, so that the owner can rationally observe the user’s needs; improve the quality of your restaurant through the analysis of reviews, so as to increase your popularity and attract more customers to come to spend; through this link, customers can also improve their own experience and experience better services.