Inner Mongolia's Hetao Irrigation District (HID) is the largest gravity-fed irrigation district in Asia, serving as a major grain and oil production base in China. The irrigation system in HID adopts a non-growing season irrigation regime comprising Autumn Irrigation (AI) and Spring Irrigation (SI) to mitigate soil salinization before crop cultivation. However, in response to the growing scarcity of Yellow River water resources, Bayannur in Inner Mongolia has been required to reduce the AI water use from 1.6 billion m3 to 1 billion m3 in 2023. Therefore, this study investigated the spatiotemporal evolution of crop planting structure and irrigation areas in the HID and proposed an optimized water allocation scheme for AI and SI through the Simultaneous Heat and Water (SHAW) model integrated with field data. The results showed that sunflower occupied the largest planting area in the HID, followed by maize and wheat from 2016 to 2023. The planting areas of sunflower and maize exhibited an increasing trend, whereas that of wheat decreased. Moreover, the area of SI decreased while that of AI increased before 2023, whereas both reversed their respective trends in 2023. The optimized water allocation scheme recommends irrigation quotas for lightly saline-alkali land at 780-1230 m3 /ha for SI and 1650 m3 /ha for AI. For moderately saline-alkali land, the suitable irrigation quotas are 1080-1305 m3 /ha for SI and 1800 m3 /ha for AI. The recommended area of AI is 336107 ha with 1.183 billion m3 of water use, and that of SI is 383380 ha with 0.908 billion m3 . This study provided optimized irrigation strategies under water constraints, contributing to improve water use efficiency, reduce soil salinization, and sustainable agricultural development in cold and arid region.