In　order　to　reduce　the　communication　cost　of　the　multi-agent　cooperative　control　in　obstacle　environments,　this　paper　designs　a　distributed　event-triggered　optimization　control　method.　The　method　jointly　designs　the　event-triggered　communication　mechanism　and　control　policy,　determines　the　communication　triggering　conditions,　and　reduces　redundancy　in　the　real-time　or　cyclic　communication.　Meanwhile,　in　order　to　avoid　collision　between　the　agent　with　other　agents　or　obstacles,　an　exponential　function-based　collision　penalty　term　is　designed　to　give　a　penalty　exponentially　when　the　agent　approaches　obstacles　or　other　agents.　The　method　is　applied　to　the　multi-agent　deep　deterministic　policy　gradient　(MADDPG)　algorithm　and　simulated　on　the　multi-agent　particle　environment　(MPE)　with　added　obstacles.　Simulation　results　show　that　this　method　can　complete　the　collision-free　multi-agent　cooperative　control　task　with　superior　performance　while　reducing　the　amount　of　data　transmission.　©　2024　Chinese　Academy　of　Sciences.　All　rights　reserved.