Abstract: By establishing evaluation indicators and designing a data mining method combining clustering analysis and association analysis， this paper analyzes staff research capacity at colleges and universities. The results showed that this proposed method can well reflect the problems of sampled colleges and universities including staff educational history， key talents development ability of horizontal projects and their research capacity diversity. Accordingly， some strategies were put forward， which include stressing talent introduction from key universities， strengthening academic exchanges among key talents and setting up specific encouragement policies， which could help personnel departments at colleges and universities to make scientific talent policies and improve the overall research capability.

Key words: data mining, college and university, staff evaluation, research capacity