Nowadays, the present incremental database association rule mining is an important area as data mining, has been widely used in education, medical, health and other fields, so it has become the most active data min are change and unknown. If use the Apriori algorithm to calculate, on the one hand it is difficult to achieve good results, on the other hand, a great impact on support changes we can not determine the support change. So with the mechanism of qualitative attribute theory and attribute computing network boundary learning algorithm with IUBM algorithm, we propose an algorithm for mining association rules based on incremental qualitative attribute. For example, in order to score crossed the enrollment system, qualitative datum point, it may completely change the life of a student's life. With the experiments show that, this algorithm reduces the redundant rules generated in the incremental mining association rules in large-scale data processing, at the same time the mining efficiency has been greatly improved. Apply the researches to prediction of College Students' employment for graduates to understand the learning situation of application is very meaningful.