The possibilities of cloud characteristics retrieval with multiple-field-of-view Raman lidar are considered. It has been shown that the Raman lidar return is sensitive to two cloud characteristics; the scattering coefficient and the effective droplet size. This sensitivity is studied and the optimal receiver fields-of-view (FOVs) for cloud sounding are recommended. The optimal FOV values are estimated to be approximately R/H (R, the collecting optics radius, H, the cloud altitude) to measure the scattering coefficient profiles, and approximately 0.01z/H for the droplet size measurements (z, the cloud thickness). The algorithm based on the iterative scheme and singular value decomposition as a regularization procedure is presented and verified using computer simulation. The recommendations for profile retrieval with variable altitude resolution are given.