Collaborative management of the water‐energy‐food‐ecosystems (WEFE) nexus can contribute significantly to sustainable development. However, multiple decision‐making levels with diverse preferences and multiple uncertainties in different forms pose intractable challenges to the management process. In this study, a novel optimization method named as multi‐level chance‐constrained fuzzy programming (MCFP) is developed to jointly manage the WEFE nexus. MCFP has advantages in evaluating trade‐offs among multiple competitive decision makers, solving decentralized planning problems with hierarchical structure, and tackling uncertainties expressed as randomness and vagueness. MCFP is then applied to the WEFE nexus in Central Asia, where five countries, 43 states, six water sources, and eight water users are involved over a long‐term planning horizon (2021–2050). A set of scenarios are designed to reflect decision‐making preferences based on different irrigation efficiencies, food, ecological and electricity demands as well as constraint‐violation probability and system credibility levels. The major findings are: (a) the proportion of agricultural water allocation would reduce to 45.4%–56.6% by 2050 to save more water for ensuring ecological restoration and energy supply; and (b) in order to balance water demands and support regional sustainable development, policymakers should sacrifice some of the benefits, set strict arable land limits for cereal crops, improve irrigation efficiency through adopting drip and sprinkler irrigation, and avoid the effects of the irrigation efficiency paradox. The findings are helpful for policymakers in gaining insight into the interrelationships of water, energy, food and ecosystems as well as making decisions for collaborative management of the WEFE nexus system.