Accurately　representing　the　joint　probability　density　of　wind　vector　is　significant　for　wind　energy　assessment,as　well　as　structural　design　of　wind　turbine.　This　study　will　focus　on　improving　the　accuracy　of　describing　joint　probability　density　of　wind　vector.　Firstly,the　wind　vector　is　represented　as　orthogonal　components,　a　mixed　Normal　distribution　is　used　to　represent　their　marginal　probability　density,and　a　method　for　estimating　the　initial　values　of　parameters　is　proposed.　Then,　based　on　Copula　function,the　correlation　between　the　orthogonal　components　of　wind　speed　is　considered　to　obtain　their　joint　probability　density,　the　Copula　parameter　is　estimated　by　the　Feast　Squares　estimation,　and　then　joint　probability　density　of　wind　vector　is　obtained　through　Jacobian　transformation.　Finally,　the　hourly　average　wind　speed　and　direction　data　in　point　S3　and　S7　from　the　Indian　Wind　Energy　Research　Institute　is　used　to　compare　with　previous　joint　probability　density　results　of　wind　vector　obtained　based　on　the　Copula　function　and　wind　speed　and　direction　variables.　The　results　show　that　the　joint　probability　density　of　wind　vector　obtained　by　proposed　method　is　more　accuracy　than　previous　method.　This　study　can　provide　a　theoretical　basis　for　the　exploitation　and　utilization　of　wind　energy.　©　2024　Editorial　Office　of　Chinese　Journal　of　Computational　Mechanics.　All　rights　reserved.