A　wavelet　estimation　method　is　presented　herein　to　estimate　deep　pit　settlement.　In　this　method,　the　pit　settlement　is　decomposed　into　the　trend　settlement　and　the　stochastic　settlement　by　using　Wavelet　Analysis　based　on　the　characteristic　of　influencing　factor.　The　model　identifier　is　established　by　using　artificial　neural　network　(ANN),　and　trained　to　approximate　the　trend　settlement.　Then,　the　prediction　controller　developed　could　be　applied　for　estimating　the　actual　settlement.　Finally,　the　verification　examples　show　that　the　WIAN　is　an　effective　tool　for　predicting　the　pit　settlement　dynamically,　high　precision　could　be　expected　and　achieved.　©　(2012)　Trans　Tech　Publications,　Switzerland.