In the context of the new era, the education concept has been transformed, and practical parenting has attracted much attention, and the talent cultivation mechanism of colleges and universities therefore urgently needs to be innovated to meet the requirements of the development of the times. In this paper, we optimize the original objective function of the sample of students’ civic education, use Taylor’s formula to expand the objective function approximation, calculate the complexity of the tree model, and complete the construction of XGBoost student learning behavior. Improve the K-means algorithm based on Canopy, combined with the improved SVM algorithm, to provide early warning for students’ academic performance in Civics. The assessment framework for teaching effect is designed to assess students’ Civics learning process. The results show that after the intervention of the model proposed in this paper, the students’ Civics performance improved from 60.59 to 65.63. The cognitive ability of Civics has a significant effect on social learning and individual learning behaviors at the same time, with the effects of 0.8623 and 0.9782. The innovative nurturing model of Civics in colleges and universities is improved by the proposed model in this paper.