College enrollment has skyrocketed, leading to a dramatic increase in the number of students enrolled. There is a widening chasm between the supply of talented innovators and the demand from society. Hence, it is critical to think about and do something about how colleges might assess students’ innovative capacities in an objective way and enhance the standard of talent innovation development. The evaluation of students’ innovation quality is essentially an assessment of their various innovative abilities, in order to assess the level of their innovation abilities in various aspects. The evaluation criteria are often set based on different educational goals of each university and may be adjusted according to policy orientation and social needs. The evaluation results can serve as a benchmark for the quality assessment of universities’ talent innovation cultivation and also serve the selection of innovative talents. This article utilizes an information system to obtain massive student data and employs data mining techniques to extract deeper and valuable information, which provides services for the cultivation of innovative talents in universities. Considering the low efficiency of the K-Means algorithm, this article improves its execution efficiency. Specifically, this article leverages the powerful parallel computing and massive data analysis capabilities of the Hadoop platform to parallelize the K-Means clustering algorithm. This can improve the execution efficiency and perform calculations on real student data.