In this article we consider a stratified multistage sampling in which the first stage units are selected by stratified random sampling without replacement and second-stage units from each selected unit similarly chosen. Estimates for domain of study are given. The analysis is illustrated using the data collected in an application of the design.