A new scheme of implementing wavelet transform using switched-current filters is proposed in this paper. Based on approximation theory of network function, the basic wavelet function is synthesized with the sum of rational fraction obtained by a systematic algorithm. And then the basic wavelet function filter is implemented by a parallel structure of switched-current circuits. Herein, the first derivative of Gaussian function is selected as an example with the construction procedure of approximation network and structure of switched-current filters elaborated. The filter network employs six biquadratic switched-current filters in a parallel connection, in which the S2I memory cell is used as the core of the filter. Theoretical analysis and simulation result show that the proposed scheme is superior to the original method in approximation precision, system stability and circuit performance.