Cloud　service　centers　have　the　problems　such　as　untrusted　virtual　machines　and　frequent　untrusted　migration　caused　by　instantaneous　peak.　In　this　paper,　the　trusted　virtual　machine　is　measured　and　the　dynamic　exponential　smoothing　prediction　algorithm　is　used　to　filter　the　migration　requests,　to　ensure　that　a　trusted　virtual　machine　will　migrate　at　the　trusted　migration　time.　The　trustworthiness　of　the　service　level　agreement　(SLA)　for　cloud　services　is　maintained,　and　the　efficiency　of　the　utilization　of　cloud　service　center　resources　is　improved.　©　Published　under　licence　by　IOP　Publishing　Ltd.