It　is　helpful　to　install　reactive　compensation　equipment　at　wind　turbine　terminals　to　improve　the　bad　influence　of　wind　power　on　the　voltage　quality　of　power　system.　Because　of　the　random　feature　of　wind　power　and　change　variation　of　the　load　and　system＇s　voltage,　a　mathematical　model　based　on　chance　constrained　programming　(CCP)　is　given.　And　an　algorithm　of　stochastic　simulation＇s　particle　swarm　optimization　is　put　forward　to　figure　out　the　reactive　power　compensation　capacity.　The　model＇s　objective　function　is　to　minimize　cost　under　consideration　of　probability　distribution　of　wind　speed,　the　characteristic　of　active　and　reactive　power　of　wind　turbines,　the　way　of　power　flow　calculation　in　wind　power　integrated　system,　and　so　on.　By　analyzing　the　variation　of　reactive　power,　voltage　and　power　factor　of　wind　farm　and　utilizing　the　given　way　to　optimize　reactive　power　compensation　in　a　distribution　network,　the　model　and　algorithm　are　proved　effective.