In the novel system for off-line signature verification, both static and pseudodynamic features are extracted as original signal, which is processed by Discrete Wavelet Transform (DWT) for the purpose of enhancing the difference in time domain between a genuine signature and its forgery. The verification scheme is achieved by combining the proposed fuzzy net output in each sub-band level. In addition, how to train the fuzzy net and get the appropriate training samples are also introduced. The experimental results with English and Chinese signature databases confirm the effectiveness of the proposed technique and show its ability to yield high verification rates.