Abstract:[ Purpose/Significance] Management Science has accumulated a wealth of research results. Based on the exsisting literature in this field, multidimensional analysis of research topics can help to reveal research features and evolution of research topics among different universities. [ Method/Process] This article takes papers from twenty-two A-class important journals in Management Science, from 2005 to 2014, as the data source, with adopting method of LDA (Latent Dirichlet Allocation) topic modeling and cluster analysis. By mining the content, theme and feature of papers that published by thirty-nine universities of the 985 project in different time period, the study illustrates the research topics and evolution in the field of Management Science in universities, from multiple dimensions such as school, journal and time. [Result/Conclusion] The result shows that, some research themes are relatively common in the field of Management Science in China, while different universities also have their own characteristic in the topics of their studies; the same subject occupies differently in position and weight in the development trend of different universities; based on the similarities of research topics, four groups of universities that have different characteristic research topics in Management Science have formed as result.