The　wavelet　transform　is　used　as　time-frequency　representation　for　the　modal　identification　from　operational　vibration　measurements.　In　this　paper,　a　wavelet　transform　identification　technique,　which　based　on　covariance-driven　and　Singular　Value　Decomposition　(SVD),　is　discussed.　The　technique　starts　with　the　calculation　of　covariance　from　random　response　of　the　system　that　converted　into　the　time-scale　domain　using　a　continuous　wavelet　transform.　It　can　be　shown　that　the　ridges　of　the　covariance　wavelet　coefficient　magnitudes　contain　modal　information　of　the　system.　By　taking　SVD　of　the　covariance　matrix,　the　ridges　of　the　covariance　wavelet　coefficient　are　extracted　and　be　used　to　estimate　modal　parameter　of　the　system.　A　real　bridge　under　ambient　excitations　is　performed　to　demonstrate　the　proposed　technique.　The　results　obtained　are　comparable　with　those　previously　obtained　from　the　peak　pick　method　in　frequency　domain　and　stochastic　subspace　identification　in　time　domain.　The　proposed　technique　is　proved　reliable　and　efficient,　which　can　be　used　to　identify　the　modal　information　of　a　full-size　bridge　under　operational　conditions.