An　improved　dynamic　wavelet　fingerprint　(DWFP)　technique　was　developed　to　characterize　nonlinear　ultrasonic　effects.　The　white　area　in　the　fingerprint　was　used　as　the　nonlinear　feature　to　quantify　the　degree　of　damage.　The　performance　of　different　wavelet　functions,　the　effect　of　scale　factor　and　white　subslice　ratio　on　the　nonlinear　feature　extraction　were　investigated,　and　the　optimal　wavelet　function,　scale　factor　and　white　subslice　ratio　for　maximum　damage　sensitivity　were　determined.　The　proposed　DWFP　method　was　applied　to　the　analysis　of　experimental　signals　obtained　from　nonlinear　ultrasonic　harmonic　and　wave-mixing　experiments.　It　was　demonstrated　that　the　proposed　DWFP　method　can　be　used　to　effectively　extract　nonlinear　features　from　the　experimental　signals.　Moreover,　the　proposed　nonlinear　fingerprint　coefficient　was　sensitive　to　micro　cracks　and　correlated　well　with　the　degree　of　damage.　(C)　2016　Elsevier　Ltd.　All　rights　reserved.