Methods for risk analysis of flood levee systems that account for hydraulic and hydrologic uncertainties are incorporated into optimal design procedures. The uncertainties are considered in the expected damage functions so that the objective is to determine the optimal flood levee system design that minimizes the total expected cost. Dynamic programing (DP) and discrete differential dynamic programing (DDDP) models are developed for the optimization. The DDDP approach tremendously reduces the computation effort and provides more accurate results when compared to the DP. Through the use of an example, different design philosophies (level of risk consideration) are examined.