This　paper　studies　multivariate　wavelet　regression　estimators　with　errors-in-variables　under　strong　mixing　data.　We　firstly　prove　the　strong　consistency　for　non-oscillating　and　Fourier-oscillating　noises.　Then,　a　convergence　rate　is　provided　for　non-oscillating　noises,　when　an　estimated　function　has　some　smoothness.　Finally,　the　consistency　and　convergence　rate　are　discussed　for　a　practical　wavelet　estimator.