Recent　advances　in　edge　computing　had　significant　impacts　on　the　development　of　internet　of　vehicles.　However,　the　internet　of　vehicles　are　confronted　with　serious　challenge　in　terms　of　deployment　unbalanced,　computing　resource　allocation　unbalanced　and　time　delay　on　real-time　computing　because　of　the　roadside　edge　computing　nodes.　In　this　paper,　a　task　queuing　model　and　computing　resource　dynamic　scheduling　policy　are　proposed.　Firstly,　a　mobile　edge　computing　(MEC)　architecture　is　designed　to　build　elastic　distributed　services　in　vehicular　networks　environment.　Then,　the　M/M/1　task　queueing　model　is　constructed　by　queue　theory　and　tasks　generated　by　virtual　vehicles　are　assigned　and　changed　access　to　the　MEC　server　according　to　cell　switch　model.　Moreover,　the　additive　increase　multiplicative　decrease　(AIMD)　algorithm　are　presented　to　solve　the　imbalance　of　computing　resource　allocation.　The　simulation　results　illustrate　the　effectiveness　of　the　proposed　method.