Abstract: In order to solve the problem of existing intelligent compaction indicators not considering the detection value attribute data, the AHP method and grey correlation degree method were adopted, comprehensively considering the roller working parameters, spatial position and other attribute data contained in the detection values. An AHP grey correlation degree model for intelligent compaction detection values was established, and the rationality of the model was verified through calculation. Based on the on-site intelligent compaction test data, it is proposed to use the optimal compaction detection value M_R solved by the model as the compaction representative value. M_R and the current intelligent compaction indicator M_V are respectively compared with the traditional compaction indicators E_vd and K₃₀ for correlation verification and indicator dispersion analysis. The results show that the correlation verification results of M_R, M_V with E_vd and K₃₀ are all greater than 0.7, and the accuracy of M_R evaluation results is better than M_V. In addition, the overall variability of M_R indicators is also less than M_V, so M_R indicators can more accurately reflect the true compaction situation of the compaction unit. This study has reference value for the improvement of intelligent compaction detection indicators.