Embodiments of the invention provide a complex PDF structure analysis method and device based on a neural network. The method comprises the steps of obtaining feature information of a PDF document; carrying out coarse particle division on the feature information of the PDF document according to a maximum entropy model to obtain a layered paragraph of the PDF document; converting layered paragraphsof the PDF document according to a two-layer bidirectional language model trained in a large-scale corpus set to obtain paragraph word vectors, and compressing the paragraph word vectors to obtain paragraph semantic vectors; and inputting the paragraph semantic vector into a multi-layer bidirectional long-short-term memory network to obtain a hierarchical sequence of all paragraphs of the PDF document. The technical problem that the generalization ability is poor due to the fact that the PDF document structure is not single is solved, and the technical effects that the limitation of manual design rule logic is avoided, the complex PDF document structure can be analyzed at a high level, and the generalization ability is high are achieved.