Abstract: After collecting 142 sample datasets from a topic sentence natural language corpus, lexical category sequencing of natural language is constructed by a Chinese natural language processing platform. The lexical category sequence classification model is composed of noun, verb, adjective and pronoun after the coarse graining processing of language structure. The results show that the lexical category sequence classification model based on the lexical category content do archive the accuracy of 90%, and the same model based on the position of word order do archive the accuracy of 95%. Thus, it proves that the lexical category sequence of natural language classification model would be valuable in linguistic cognitive. It not only reveals the relationship between language and psychology Information, but also assesses the implicit psychology information scientifically through the objective linguistic notations.