In　view　of　the　deficiency　of　previous　researches,　this　paper　draws　on　the　idea　of　＇decompositionintegration＇　and　combines　the　maximal　overlapping　discrete　wavelet　transform　method　with　the　high-order　moment　portfolio　framework　based　on　the　time-frequency　domain　perspective　to　construct　a　waveletintegrated　higher-order　moment　portfolio　model.　Since　the　wavelet　functions　are　various　with　different　properties,　this　paper　proposes　the　idea　of　＇decomposition-mixing-integration＇　and　constructs　a　hybrid　wavelet-integrated　higher-order　moment　portfolio　model.　Based　on　data　of　stock,　crude　oil　and　gold　markets,　this　paper　adopts　the　constructed　model　to　build　portfolios,　and　conducts　out-of-sample　testing　and　analysis　on　them,　after　which　stability　testing　was　carried　out.　The　empirical　results　show　that,　first,　compared　with　the　control　group,　the　wavelet-integrated　high-order　moment　strategy　achieves　a　significantly　better　investment　effect,　bringing　higher　returns,　Sharpe　ratio　and　utility　level　to　investors.　Second,　by　combining　the　advantages　of　many　wavelet　functions,　the　hybrid　wavelet-integrated　highorder　moment　strategy　achieves　better　effects　and　significantly　optimizes　the　wavelet-integrated　high-order　moment　strategy.　The　stability　test　verifies　the　above　conclusions.　©　2020,　Editorial　Board　of　Journal　of　Systems　Engineering　Society　of　China.　All　right　reserved.