Asphalt　mixture　is　a　typical　granular　material,　and　its　macroscopic　properties　are　closely　related　to　the　interaction　behavior　of　aggregate　particles　at　the　micro　scales.　Discrete　element　method　(DEM),　as　an　important　numerical　simulation　method　in　granular　mechanics,　plays　an　increasingly　important　role　in　analyzing　the　aggregate　packing　characteristics,　skeleton　features,　predicting　mechanical　and　pavement　performance　of　asphalt　mixtures.　However,　the　complexity　of　the　composition　of　asphalt　mixtures,　including　the　particle　size,　shape,　spatial　distribution　of　aggregate　particles,　and　the　viscoelasticity　of　asphalt　binders　sensitive　to　temperature,　load,　and　time,　increases　the　difficulty　of　constructing　DEM　models　of　asphalt　mixtures,　selecting　contact　constitutive　models,　and　determining　parameters,　which　directly　affects　the　accuracy　and　efficiency　of　DEM　simulations.　In　this　review,　the　generation　methods　of　idealized　models,　image　models,　and　user-defined　models　currently　used　to　generate　DEM　models　of　asphalt　mixtures　were　summarized　in　detail.　Conventional　contact　constitutive　models,　including　elastic　and　viscoelastic　contact　stiffness　constitutive　models,　as　well　as　fracture　models　including　bonding　models　and　cohesive　softening　models,　were　discussed.　Moreover,　special　contact　constitutive　models,　such　as　damage,　self-healing,　and　two-stage　contact　models　considering　compaction　evolution,　were　also　summarized　and　discussed.　Then,　the　contact　constitutive　model　parameters　were　classified　and　related　determination　methods　were　summarized,　and　the　latest　application　directions　of　DEM　in　current　asphalt　mixture　research　were　classified.　On　this　basis,　relevant　discussions　were　carried　out,　and　the　challenges　and　future　development　prospects　faced　by　DEM　in　asphalt　mixture　research　were　proposed.