The use of big data to build a “smart party building” model, and do a good job of related security, is conducive to more quickly control the overall situation of party building work. In this paper, we propose a method for determining party members’ rank by combining principal component analysis and K-means clustering, and combine the improved decision tree algorithm with big data for party building to realize the design of a decision support system for intelligent party building. The research results show that the proposed joint analysis method of dimensionality reduction and clustering can classify party members into different grades, and its k-s test significance is 0.214>0.05, obeys normal distribution, and the observed values are more in line with the expected values, so it can effectively assess the ideological dynamics of party members. The precision and accuracy of the designed decision support system are more stable, maintaining at about 87% and 94% respectively, which is significantly better than the traditional system, and is of great significance to improve the efficiency of management of party members’ ideological dynamics in colleges and universities.
