The invention discloses a journal influence evaluating method based on academic big data. According to the method, Average influence of all theses in every journal is used for representing the influence of the journal. A method for calculating the journal influence is a PageRank-based algorithm, namely, AR-PJageRank, the algorithm takes a similarity, an author H index and other factors between a thesis and another thesis referring to the thesis into consideration, AR makes the algorithm be a weighting algorithm, and in the algorithm iteration process, when all the theses vote a thesis cited by the theses, different contributions can be given due to the similarities between the thesis and the referenced thesis; PageRank makes the importance of each thesis greatly increased, an H index of each thesis is calculated through the sum of the author H indexes, and the H index serves as a label of the thesis so that difference of each thesis can gradually emerge in the calculation process. Accordingly, the method can accurately distinguish the influence of each thesis, and finally the influence of each thesis is obtained.