Abstract: Evaluation on compaction uniformity is a key method for preventing uneven settlement in earth-rock dams after construction.A full-section assessment of compaction quality provides essential data for evaluating compaction uniformity.Under current specifications, traditional sampling methods may lead to deviations in full-section estimation and uniformity evaluation due to limited sample sizes.To address this, this paper proposes a rapid evaluation method for assessing the compaction uniformity of earth-rock dam materials based on small-sample detection.Firstly, the Kriging interpolation method (KRG) was used to generate the distribution cloud map of compaction quality, enabling full-section estimation of compaction quality.Then, based on the distribution cloud map, the spatial aggregation characteristics of unqualified weak areas were analyzed.On this basis, considering the multi-source information conflict resolution mechanism, a comprehensive evaluation model of compaction uniformity based on D-S evidence theory was constructed.Engineering case studies demonstrated that the cross-validation error of this evaluation method remains within a controllable range.After multi-source information fusion, the model can effectively identify the spatial aggregation state of unqualified weak areas and enables rapid evaluation of compaction uniformity.This approach provided a basis for guiding construction quality control.