The　asset　allocation　strategies　don＇t　take　into　account　correlation　risks　in　classical　finance　theories,　but　the　correlation　of　markets　and　assets　change　all　the　time　in　nature,　which　raise　the　risks　of　portfolio.　Unlike　portfolio　choice　of　traditional　finance　theories　keeping　a　complete　market　setup,　we　reproduce　the　asymmetric　tail　dependence　of　capital　market　with　symmetric　Joe-Clayton　Copula　and　solve　the　portfolio　efficient　frontiers　and　strategies　when　correlation　risks　exist　with　CVaR　technique.　Through　the　empirical　study　in　Shanghai　and　Hong　Kong　market,　we　discover　that　ignoring　upper　tail　or　lower　tail　dependence　will　affect　investors　to　estimate　the　risks　of　portfolio,　and　will　make　the　portfolio　suffer　extreme　negative　returns;　quantify　and　control　the　tail　dependence　will　improve　the　performance　of　portfolio.