To improve the control accuracy of the multiple-regression feedback model of intelligent compaction control technology based on compaction parameters and filler properties, this study constructed the multiple regression model considering only the compaction parameters, and then optimized the feedback model with the moisture content as the correction parameter on the basis of analyzing its influence law on the observation value of vibrations. The results of previous feedback model and optimized feedback model show that although the intelligent compaction construction control using these two feedback models meet the requirements of specification, the qualified rate of the optimized feedback model is greater than the previous feedback model and the control accuracy is improved by 54.1%, which proves the scientificity and feasibility of the optimized feedback model.