Design and optimization of steam power systems (SPS) should not only optimize the economic target of the system but also ensure the safe and stable operation of the production process against complex uncertainties. For the uncertainty and complexity existing in SPS, the study proposes an interval two-stage stochastic programming (ITSP) model by merging two-stage stochastic programming (TSP) and interval-parameter programming (IPP). The approach has the advantages that uncertainties are represented by discrete random variables, interval numbers, and their combination and can be effectively communicated into the optimization processes and resulting solutions; in addition, it incorporates the regulated equipment selection policies directly into its optimization process and can put forward corresponding scheduling and adjustment strategies for various policy scenarios caused by uncertainties. The optimal scheme can be obtained as stable interval solutions and provides ample decision-making space for decision makers, which achieves the comprehensive optimum of economy and flexibility.