To handle eco-environment vulnerability, water shortage and complex uncertainties in irrigation districts in arid areas of Northwest China, taking Hongyashan Irrigation District in Shiyang River Basin as an example, a type-2 fuzzy multi-objective programming model for optimization of agricultural-ecological water and land resources allocation was developed. The developed model integrated type-2 fuzzy sets, fuzzy credibility-constrained programming, and multi-objective programming, with the objectives of minimum irrigation loss, maximum ecological vegetation irrigation satisfaction, minimum ecological vegetation irrigation cost and maximum economic benefit of main grain crops. The decision variables were surface water, groundwater and planting structure among 10 decision-making units in Hongyashan Irrigation District. The optimal allocation schemes of irrigation water and planting structure under different credibility levels and uncertainty degrees were obtained by solving the model. Optimal results showed that the model coupled with type-2 fuzzy sets can provide abundant allocation schemes. The sensitivity of allocated water to credibility levels was higher than uncertainty degrees, and the planting structure was not sensitive to credibility levels. When the uncertainty degree was 0.5 and the credibility level was 0.7, the ecological vegetation was totally irrigated by surface water, crops were irrigated by surface water and groundwater, and the yield and economic benefit of maize are higher than that of wheat.Compared with previous studies, the research more considered the irrigation requirement of ecological vegetation, and the optimization results were more realistic and reasonable. The research results can provide decisionmakers with allocation schemes which were in line with the actual situation of irrigation districts and provide scientific guidance for the modernization of irrigation districts in arid areas of Northwest China.