The dynamics governing high speed milling has been extensively explored for decades, both numerically and experimentally. The process loses its dynamic stability and becomes chaotic through either Neimark-Sacker or period doubling bifurcation. While its associated response is still bounded in the time domain, however, it could become unstably broadband in the frequency domain, thus causing the uneven cutting surface on the workpiece and tool damage. A discrete-wavelet-based feedforward adaptive controller is developed to stabilize system response in the time and frequency domains simultaneously. An adaptive controller along with an adaptive filter effective for on-line system identification is implemented in the wavelet domain. By controlling both time and frequency responses, the presented controller design is demonstrated to effectively suppress milling chatter and restore the system back to dynamic stability.