With　the　rapid　growth　of　urban　motor　vehicle　ownership,　formulating　reasonable　parking　price　to　alleviate　the　parking　problem　is　increasingly　important.　Based　on　parking　survey　data　in　urban　business　district,　the　multi-agent　modeling　method　is　adopted　to　build　a　simulation　environment.　Then　behavior　decision-making　rules　for　the　travelers＇　parking　process　are　built　through　programming,　and　the　simulation　platform　for　curb　parking　under　time-varying　parking　price　is　established.　Finally,　the　implementation　effect　of　time-varying　parking　price　is　analyzed.　The　findings　show　that　the　implementation　of　the　time-varying　parking　price　not　only　makes　the　utilization　of　parking　resources　more　balanced　and　reduces　the　cruising　time　for　parking,　but　also　increases　the　vehicles　speed　on　surrounding　roads.　The　study　can　provide　decision-making　suggestions　for　improving　parking　efficiency　and　alleviating　regional　traffic　congestion.　Copyright　©　2019　by　Science　Press.