The invention belongs to the field of document application and discloses a method for identifying data form in document and a device thereof. The method includes that: text in PDF document is extracted; the text is divided according to the attribute of the extracted text, thus obtaining a division result; the division result is judged and processed, and identification is carried out, thus generating a data form; the data form is stored into an independent intermediate data structure; and the data form in the intermediate data structure is restored according to target document format. In the invention, data form in PDF is accurately converted, editability after conversion is greatly improved, and manual processing on document after conversion is reduced.