Digital signal processing techniques have already proven to be powerful tools for the identification of protein-coding regions. Proper selection of sliding window length is an important factor in improving the identification accuracy in Short-Time Discrete Fourier Transform (ST-DFT) method. A novel weighted mean multi-window (with different sliding window lengths) ST-DFT with Singular Value Decomposition (ST-DFT-SVD) method for the identification of protein-coding regions is proposed in this paper. The proposed method has two modes of operation: training mode and testing mode. The performance of the proposed method is evaluated in comparison with other existing methods such as ST-DFT, anti-notch filter and multi-stage filter. The results show that the proposed method provides superior performance in terms of identification accuracy. © 2011 Inderscience Enterprises Ltd.