With the vigorous development of the Internet, an increasing number of network media and social platforms have stepped onto people's stage. These products enter the consumer's vision in a variety of forms and become part of the fashion. Compared with the traditional business model, the large-scale use of natural language website as a new field is rising rapidly, providing users with more rich, convenient and efficient services. The purpose of this paper is to study how to deal with the word-of-mouth effect of green brand in e-commerce compared with the traditional business model, and put forward solutions. By analyzing the customer behavior characteristics and purchase intention of online shopping, this paper designs a brand reputation management system, and tests the performance of this system. The test results show that the response speed of brand D is the slowest, up to 17 s, and the response speed of brand A is the fastest, 12 s. The stability of brand A is 96%, that of brand B is 98%, that of brand C is 96%, and that of brand D is 92%. Efficient response speed can improve the degree and effect of enterprises’ influence on natural language processing platform systems in large market environments, thereby improving the effectiveness of user experience and contributing to promoting green brand promotion.