The　unit　commitment　(UC)　problem　in　wind　power　integrated　system　is　not　a　traditional　certain　problem　because　of　wind　power＇s　random.　The　solution　which　is　economic　and　reliable　is　difficult　to　attain　by　traditional　ways.　This　paper　puts　forwards　a　mathematical　model　of　UC　in　wind　power　integrated　system　based　on　chance　constrained　programming　(CCP).　It　describes　the　related　constrained　conditions　by　probability　form,　and　transforms　UC　problem　into　inside　and　outside　optimization　sub-problems.　The　outside　sub-problem　is　units＇　on/off　status　optimization.　It　is　effectively　solved　by　discrete　binary　particle　swarm　optimization　(BPSO)　and　heuristic　searching　strategy.　The　inside　sub-problem　is　economic　dispatch　(ED).　Considering　the　random　of　wind　power,　the　sub-problem　is　solved　by　improved　particle　swarm　optimization　(PSO)　based　on　stochastic　simulation　which　avoids　local　optimal　solution　and　ensures　the　feasibility　of　generation　plans.　The　optimization　algorithm　is　proved　feasible　and　effective　by　testing　a　10-unit　system　and　being　compared　with　other　methods.