The invention discloses a named entity linking method fusing priori information. The method comprises the following steps of (1) extracting a character string-candidate entity table, a person name list and a place name list from a Wikipedia data dump and a Freebase data dump; (2) representing articles in the Wikipedia data dump as tf-idf features, and extracting universal features of character strings relative to candidate entities; (3) performing inquiry expansion on entity mention, and by using the character string-candidate entity table in the step (1), generating the candidate entities forthe entity mention; (4) extracting the features of the article where the entity mention is located, thereby obtaining an inverse document frequency and an important word collision rate of the article; and (5) by using the features extracted in the steps (2) and (4), calculating correlation degrees between the entity mention and the candidate entities thereof, and taking the entity with the highest correlation degree as an entity linking result. The limitation of corpus shortage is broken through; a reliable entity linking recommendation result is provided for a user; and the universal features of the entities are added with the priori information.