Abstract:Talent selection is a complex systematic project, and talent selection indexes should consider the limitations of generic bibliometric indexes that are not oriented to disciplines in practical application. This paper uses qualitative and quantitative analysis method, and uses mathematical subject classification(MSC), establishes discipline expert consulting team for qualitative analysis, and uses bibliometric method, analytic hierarchy process (AHP) and technique for order preference by similarity to an ideal solution (TOPSIS) for quantitative analysis, and adopts MATLAB and Derwent Data Analyzer on data analysis, and designs a system of talent selection bibliometric indicators, including discipline index, journal index and talent index, to acclimate certain discipline such as mathematics. Furthermore, topology research area is selected as an example, modeled and verified by AHP and TOPSIS, and CNCI-M and h-M are iterated after validation. The results proved that the discipline-oriented bibliometric index system is more in line with the researchers' cognition.