As modern clinical trials become much more complicated, there is a significant interest of the potential use of flexible clinical trial design to improve the efficiency and the chance of success of clinical studies. This presentation examines the use of flexible sample size design as a special case of flexible clinical trial design. The focus of the presentation will be on the need for flexible sample size design, the concept, the method, and more importantly the criteria and procedure to choose a better one among different candidate flexible sample size designs. Examples and simulation results will be given to illustrate the ideas.