A method of de-noising based on wavelet thresholding for coulostatically-induced transients (CITs) is presented. This filtering was illustrated with CIT and its associated impedance spectra, and comparisons have been made with two other de-noising methods: data smoothing and digital filtering. Numerical tests analyzing the performance of different decomposition levels were conducted, so the optimization of wavelet decomposition levels for CIT was also obtained in this work. The results reported here demonstrate that wavelet de-noising is applicable to CIT in the time domain, where traditional filtering can also be successful. Moreover, in cases where traditional methods fail, wavelet de-noising can remove noisy data from impedance spectra of CIT. Therefore, the proposed method shows consistently superior de-noising performance over traditional methods in both the time and frequency domains. Practically, the proposed method is effective when the decomposition levels are from five to seven. (c) 2006 Elsevier Ltd. All rights reserved.