Most advanced control applications rely on good dynamic process models. The performance of the control system depends on the accuracy of the model used. Typically, such models are developed by conducting off-line identification experiments on the process. These identification experiments often result in input-output data with small output signal-to-noise ratio, and using these data results in inaccurate model parameter estimates. Prefilters are used to separate useful information from the noise in the input-output data and to improve parameter estimates. A systematic design procedure for selecting a prefilter using discrete wavelet transforms is presented. The design procedure provides explicit information on the compromises in prefilter design, interpreted in terms of parameter variance and bias. The prefilter design procedure is then applied to identify a second-order output error model.