A new scheme of implementing Morlet complex wavelet transform using poles-shared Switched-current (SI) circuits was proposed， in which a hybrid method in time and frequency domain was presented for approximation of Morlet complex wavelet. By decomposing the Gaussian envelop, which is a component of the Morlet complex wavelet, an approximation optimization model in time domain was designed, which can be solved in universal optimization algorithms. By using the periodic characteristics of the sine and cosine signals, the Laplace transforms of the approximated Morlet complex wavelet can be simplified. The rational real and image parts of the approximated Morlet complex wavelet have shared poles because the product of sine and exponential and that of cosine and exponential have same poles in s-domain. A kind of SI complex second order section circuit was designed based on the bilinear z-transform integrator module. Then it was used to synthesize the Morlet complex wavelet base circuit. By adjusting the circuit’s switch clock frequency, the wavelet transform in other scales can be realized. The comparative analysis demonstrates that the proposed approximation method is better than the Padé transform and Maclaurin series method in accuracy and stability. Furthermore, the circuit designed has the advantages of more simple structure, lower power consumptions and smaller volumes, compared with the existing method. Simulation results verified the effectiveness of the proposed scheme.