In this study, a dual‐interval two‐stage stochastic programming (DTSP) approach is developed for planning water resources through a market‐trading‐oriented mechanism under uncertainty. The DTSP method is used for policy analysis of water trading in the Kaidu‐kongque River Basin with the efficient and sustainable manners to relieve the pressures of economic development and population growth in northwestern China. The obtained results reveals that a market‐trading‐oriented mechanism under the adjustment/regulation by water managers would be a sustainable/effective manner to allocate water resources against the market failure. Meanwhile, the trade‐off between governmental regulation and market behavior has generated a more efficient trading mode with overall consideration of factors such as food safety, drinking safety, environmental protection, and regional development in a watershed system, which could bring about a number of beneficial impacts on regional water resources development. Moreover, the DTSP method is also appropriate for discriminating varied cases associated with various levels of economic influences since the penalties could be exercised with the recourses actions against any infeasibility.