With　the　rapid　development　of　communication　technology,　the　scale　of　user　data　grows　exponentially.　In　the　diversified　service　scenarios,　the　data　processing　speed　is　required　more　strictly　in　order　to　guarantee　the　service　quality.　Mobile　edge　computing　technology　makes　the　server　closer　to　the　terminal　and　assists　the　terminal　to　process　data　nearby,　bringing　better　user　experience.　In　mobile　edge　computing　system,　most　terminal　devices　have　strong　mobility,　and　it　is　easy　to　cause　service　interruption　because　terminal　devices　are　out　of　the　service　scope　of　the　server　after　a　few　moments.　In　addition,　the　importance　of　the　offloading　task　varies　from　one　task　to　another,　so　the　importance　index　should　also　be　considered　in　the　offloading　request　scheduling.　This　paper　mainly　studies　the　request　scheduling　problem　in　the　field　of　computational　offloading,　and　divides　the　problem　into　two　stages:　two-dimensional　decision　making　and　offloading.　During　the　research,　the　mobility　of　terminal　devices　and　the　characteristics　of　tasks　are　considered,　and　a　reward　evaluation　algorithm　based　on　DDQN　(Double　Deep　Q　Network)　is　designed　to　generate　scheduling　strategies.　©　The　Author(s),　under　exclusive　license　to　Springer　Nature　Singapore　Pte　Ltd.　2024.