As　the　aging　of　asphalt　pavement　intensifies,　rejuvenator　has　attracted　much　attention　as　an　important　repair　material.　However,　the　traditional　design　method　of　rejuvenator　has　the　shortcomings　of　long　cycle,　low　efficiency,　and　unclear　molecular　mechanism,　resulting　in　a　poor　rejuvenation　effect.　Therefore,　84　molecular　models　of　rejuvenators　were　constructed　using　the　exhaustive　method,　covering　different　chemical　compositions,　polarities,　and　electronic　structures.　The　solubility　and　electronic　structure　of　these　rejuvenators　were　calculated　by　molecular　simulation,　and　a　machine　learning　dataset　was　established　based　on　this.　The　artificial　neural　network　was　used　to　predict　the　relationship　between　the　molecular　structure　of　the　rejuvenating　agent　and　the　effect　parameters,　and　the　factors　affecting　the　rejuvenation　effect　were　analyzed　by　the　random　forest　model.　The　results　show　that　amino　groups　could　improve　the　rejuvenation　effect,　and　the　No.　3　substitution　position　of　Cyclic　saturate,　No.　2　substitution　of　Naphthene　aromatic,　and　No.　2　substitution　of　Straight　saturate　were　identified　as　the　best　modification　locations　to　improve　the　rejuvenation　effect.　Based　on　the　results,　three　new　molecular　structures　of　rejuvenators　were　designed,　and　simulation　calculations　show　that　their　effects　were　significantly　better　than　those　of　the　84　previously　constructed　rejuvenators.　These　findings　suggest　that　molecular　simulation　and　machine　learning　provide　a　viable　alternative　tool　for　efficient　rejuvenator　molecular　design.