In order to meet the urgent requirement for China spaceborne cloud radar detection and accurate retrieval of cloud microphysical properties， integrating the detecting advantages of dual-frequency 94/220 GHz cloud radars， and combining scattering and attenuation properties of non-spherical ice particles， a dual-frequency retrieval algorithm of cloud microphysical parameters is proposed based on the non-spherical ice particles scattering database. By assuming the gamma particle size distribution of cloud ice and establishing relationships between dual frequency ratio and retrieval variable volume median diameter， three types have been chosen. Based on it， the optimal solutions of particle size distribution parameters are obtained after backward iteratively calculating， then the microphysical parameters of cloud could be retrieved based on particle size distribution， and the error can be calculated and analyzed. By adding system noise and random noise， the retrieval errors are compared with original data and the requirement of signal-to-noise ratio is proposed. As a result， the retrieval of non-spherical ice cloud microphysical parameters based on dual-frequency 94/220GHz and iterative backward retrieval algorithm is well consistent with simulated true data. What’s more， the signal-to-noise ratio cannot be smaller than 11.39dB in order to get 30% retrieval error for ice water content under 0.2g/m³， which gives some references for domestic spaceborne cloud radar design and retrieval research of non-spherical ice cloud microphysical parameters.