The invention relates to a space-time trajectory-based user recognition method and a space-time trajectory-based user recognition device. The method comprises the steps of obtaining an interest regionof a to-be-identified user; constructing a space-time relationship network of the to-be-identified user and the user in the database; calculating the feature vector of the user to be identified and the feature vector of the user in the database by using a node2vec algorithm; calculating the similarity between the to-be-identified user feature vector and the user feature vector in the database; and identifying the user most similar to the to-be-identified user in the database through sorting of the similarities. According to the technical scheme provided by the invention, the relationship characteristics between the user behaviors and the users are comprehensively considered, and the characteristic description of the stability and difference of the user behaviors is considered, wherein theacquisition of the region of interest of the user gives more consideration to the behavior stability, the design of the network edge weight gives more consideration to the difference, and the adoption of multi-view autonomous representation learning is helpful for obtaining more robust user vector representation.