The　number　and　scale　of　data　centers　hosting　the　cloud　computing　industry　have　increased　rapidly　in　recent　years,　resulting　in　huge　power　consumption.　Therefore,　energy　conservation　and　emission　reduction　of　data　centers　is　imminent.　This　paper　proposes　an　energy-saving　scheduling　strategy,　the　core　idea　of　which　is　to　use　admission　control　and　priority　control　to　screen　and　sort　queue　tasks,　and　then　schedule　tasks　online　based　on　the　deep　deterministic　policy　gradient　(DDPG)　algorithm　to　adapt　to　the　highly　dynamic　nature　of　cloud　computing　loads　and　minimize　energy　consumption.　Simulation　experiments　confirm　that　the　scheduling　scheme　adopted　in　this　paper　is　effective　in　reducing　the　energy　consumption　of　cloud　data　centers　and　reducing　the　response　time　of　tasks.　©　2023　Inst.　of　Scientific　and　Technical　Information　of　China.　All　rights　reserved.