A new retrieval scheme is developed to infer tropical cloud properties and vertical structure, including liquid and ice water content, cloud top and base, and cloud layering. The retrieval scheme utilizes a cloud classification scheme that uses both International Satellite Cloud Climatology Project (ISCCP) cloud top temperature and a microwave index from the special sensor microwave/imager (SSM/I). Different cloud classes have different allowed numbers of cloud layers. The retrieval scheme also incorporates findings from observational studies. When multiple liquid layers are judged to be possible, a cloudiness likelihood parameter is used to determine the priority for the presence of liquid layer at each level, based on the European Centre for Medium-Range Weather Forecasts analyzed relative humidity field. Cloud liquid water path is determined using a microwave satellite retrieval. In case of multiple liquid layers, the liquid water path is partitioned and assigned to each liquid layer in proportion to the adiabatic liquid water path in each layer. Depending on the cloud class, ice water paths are determined using one of the following methods: (1) ISCCP ice optical depth; (2) a microwave ice retrieval that uses ice scattering signals at high SSM/I frequencies; and (3) a residual method that infers ice from the difference between a virtual liquid water path derived from ISCCP total optical depth and the true liquid water path determined from SSM/I. The retrieved cloud layering is indirectly validated using cloud cooccurrence climatology from surface observations. The cloud base retrieval is compared with lidar measurements obtained during the Tropical Ocean-Global Atmosphere Coupled Ocean-Atmosphere Response Experiment. ¿ 1997 American Geophysical Union