Through the tracking and analysis of the Twitter and Mendeley data of the research article, this paper probes into the characteristics and laws of network dissemination of research article. The Twitter and Mendeley data of 136 research articles published in November 2016 issue on five different types of journals was tracked and recorded. The cumulative trends and distribution of the two indicators were explored; The correlation between two indicators data recorded at different stages and between two indicators data and citation data counted by Scopus were calculated. Twitter and Mendeley cumulative value of research article shows independence to a certain degree. This means that the articles, whether in form of open access and published by different publishers, can be compared by the two indicators. In addition.the distribution characteristics of the two indicators of different stages are consistent, which means that the two indicators have early prediction function that can used to identify high-impact research articles earlier.