Unmanned　aerial　vehicles　(UAVs)　have　attracted　widespread　attention　due　to　their　high　mobility　and　strong　environmental　adaptability,　and　have　been　used　in　military　and　civilian　fields.　This　work　studies　the　method　of　using　the　deep　deterministic　policy　gradient　(DDPG)　algorithm　to　achieve　UAV　path　planning　in　complicated　environment.　Firstly,　establish　a　three-dimensional　scene　model　and　divide　the　drone　mission　process　into　three　stages:　flight,　waiting,　and　communication.　Secondly,　a　three-dimensional　deviation　degree　is　proposed　to　indicate　the　relative　position　of　the　drone,　obstacles　and　target　users,　so　as　to　improve　the　flight　performance　of　the　drone　and　effectiveness　of　obstacle　avoidance.　Finally,　the　deep　deterministic　policy　gradient　algorithm　is　used　to　plan　the　continuous　flight　movements　of　the　UAV　to　reduce　energy　consumption　and　improve　the　quality　of　service　(QoS),　while　avoiding　obstacles　and　completing　data　transmission　to　users.　The　simulation　experimental　results　show　that　the　proposed　scheme　is　effective　under　various　parameter　configurations,　and　it　is　better　than　existing　algorithms.　©　2022　Inst.　of　Scientific　and　Technical　Information　of　China.　All　rights　reserved.