The determination of the cloud parameters in atmosphere is of importance to the study of atmospheric physics and climate. These parameters are difficult to be derived directly in actual measurement and should only be derived after some related measurable physical factors are retrieved. Through analysis of space-based cloud observation, it is found that atmosphere path length, cloud-top temperature or pressure and the micro composition of different clouds can be used in the remote sensing and retrieval of cloud-top height. Some references show that the practical methods can be divided into two classes: the method using the geometry of imaging and the one using spectral radiation characteristics. The former relies on the space orientation of the light source, target and sensor controlled precisely during imaging and the latter relies on the quantitative radiation values of ground objects. When combined with the multispectral or hyper-spectral technology, they can derive abundant cloud information. Eight methods are presented.