This study is devoted to the development of a semianalytical algorithm for the determination of the optical thickness, the liquid water path, and the effective size of droplets from spectral measurements of the intensity of light reflected from water clouds with large optical thickness. The probability of photon absorption by droplets in the visible and near-infrared spectral regions is low. This allows us to simplify and modify well-known asymptotic equations of the radiative transfer theory, taking into account the fact that the single scattering albedo is close to one. Modified asymptotic equations are used to develop the inverse algorithm. We also avoid the use of the Mie theory, applying parametrizations and geometrical optics results with account for wave corrections. The main advantage of the method proposed lays in the fact that the equations derived not only provide a valuable alternative to the numerical radiative transfer solution but also are much more simple than equations of a conventional asymptotic theory. This simplicity allows both the simplification of the cloud retrieval algorithm and, even more important, the insight into various factors involved in a retrieval scheme.