Disease description texts are analyzed to reach a deeper understanding of the current status of online consultations for sleep disorders and the thematic characteristics of users with sleep disorders.Method/Process Data about sleep disorders from “haodf.com” website is collected by using a web crawler. Furthermore, the main themes about patients’ description are identified by the latent Dirichlet allocation(LDA) model.Result/Conclusion The departments of sleep disorders are more dispersed, and the main treatment is drug therapy. Online consultations could improve 83.2% of patients’ condition. The themes of patients’s disease descriptions include medication and consultation, external environment, description of symptoms, surrogate questions and causes. It is suggested that the platform and doctors should pay attention to the prognosis of patients’ medication and mental health status, and pay attention to the popularization of comorbidities.