The invention discloses a synonym extraction method based on remote supervision, and belongs to the technical field of natural language processing. The method comprises the following steps: establishing a vocabulary syntax mode model of synonyms in the field; constructing a remote supervision neural network learning model based on LSTM and CRF, and training by using domain entries to obtain a sentence sequence annotation set discovered by synonyms; and according to the annotation set, annotating and pairing candidate entities in statements in the corpus, and extracting the entities to obtain synonyms. According to the method, the corresponding vocabularies of the domain synonyms are combined by utilizing the entry characteristics based on the encyclopedic knowledge base; In the syntax mode, domain synonyms are obtained through remote supervised learning and machine autonomous learning, and the method takes machine processing as a main part and manual processing as an auxiliary part, sothat the efficiency of obtaining the synonyms is improved, and the labor cost is greatly reduced under the condition that the precision is not reduced. New words can be found through regular entry learning of online encyclopedia and analysis of hidden synonyms.