Intelligent compaction is a new road construction quality monitoring and management technology that has been widely used in the asphalt pavement compaction process. However, the lack of understanding of the compaction process and the interaction mechanism between asphalt mixture particles, which leads to over-compaction and non-uniform compaction, which seriously affects the road construction quality and service performance. In this study, the variation of asphalt mixture compaction with the number of compaction times was investigated by using a ground statistical model semi-variance function, and the effect of temperature on the compactness and its uniformity was discussed. It was found that compaction degree and uniformity did not always correlate positively with the number of passes, and that excessive compaction led to a reduction in compaction degree and uniformity. In addition, changes in compaction temperature in the field can cause significant differences in the compaction of the mixes. Higher temperature asphalt mixes have a faster increase in compaction and higher final compaction uniformity during compaction. The study shows that the dynamic response within the mix directly reflects the compaction of the asphalt mix. It is found that temperature variation does not always show a positive correlation with the compaction effect, and that the asphalt mix can only achieve the best compaction effect within the appropriate temperature range.