The invention relates to the technical field of intelligent customer service, and discloses an intelligent customer service question-answering system based on a naive Bayesian classification algorithm, which comprises the following steps: 1) the voice input is performed by a user; 2) the voice recognition module recognizes the voice; 3) the language recognition module understands the natural language through a recognition algorithm; and 4) the understood questions are input into a question and answer engine. The method has the advantages that through voice input, voice recognition, natural language understanding and question and answer engine inputting, natural language understanding comprises user intention understanding and user question quick positioning, meaning to be expressed by language of human beings in various environments can be accurately and effectively recognized, and user experience is improved. Based on artificial intelligence technologies such as a deep neural networkand syntax dependence, the intelligent robot can accurately understand user problems and has intelligent capabilities such as intelligent questioning, semantic understanding accuracy exceeds 95%, andservice quality and efficiency are improved.