Cognitive　radio　(CR)　as　one　of　the　most　promising　technologies　to　solve　the　problem　of　exhausting　of　spectrum　resource,　draws　a　lot　of　attention　recently.　However,　for　the　highly　intelligent　wireless　devices　in　the　future,　the　assumption　＇CR　devices　introduce　no　interference　to　the　primary　users　(PUs)＇　in　traditional　CR　systems　may　do　not　exist　anymore,　due　to　that　the　devices　are　always　trying　to　maximize　their　own　reward,　without　considering　the　interference　to　the　others.　In　this　paper,　the　concept　of　proactively-　optimization　CR　network　is　introduced　to　deal　with　this　problem　by　enforcing　＇punishment＇　or　＇penalty＇　to　illegal　CR　transmissions.　Under　the　framework　of　proactively-optimization　CR　networks,　with　the　reward　function　defined　in　this　paper,　the　spectrum　resource　allocation　decision　making　problem　of　primary　users　is　modeled　as　a　Markov　decision　process　(MDP),　and　an　optimal　resource　allocation　scheme　with　the　objective　of　maximizing　the　PU＇s　reward　is　proposed.　Furthermore,　extensive　simulation　results　show　that　the　proposed　scheme　improves　the　reward　significantly　compared　to　the　existing　scheme.　©　2011　IEEE.