Scientific citation measurement is an important tool to assess a scholar’s academic and research achievement. The research method based on the PageRank algorithm has two major drawbacks: (1) The influence of a paper may vary along with time. (2) There are disparities in citations of different subjects. These are not taken into consideration. Aiming at offsetting these deficiencies, with statistical approach and the introduction of time factor and subject factor, a new evaluation model based on the PageRank algorithm for the influence of academic paper is therefore established. This new model is able to effectively reveal the difference of outcomes when the time and subject factors change and measure literature quality in a more timely, comprehensive and objective manner.