[Objective] The water consumption of natural vegetation in the lower reaches of the Tarim River was analyzed in order to provide scientific guidance for ecological restoration.[Methods] Based on Landsat series images and daily meteorological data, the temporal and spatial evolution trend of vegetation water consumption from the Daxihaizi Reservoir to the Alagan section was analyzed by using the improved Penman-Monteith formula method at 30 m spatial resolution.[Results] ① The spatial and temporal distributions of natural vegetation water consumption in the research area had obvious gradient differences. The high water consumption area was mainly located along the main river and the natural overflow area. The water consumption of vegetation was generally low in the lower reaches of the Tarim Rive, and it was mainly concentrated during the middle stage of vegetation growth. ② The average water consumption of natural vegetation was 1.90×10⁸ m³/yr and increased at a rate of 2.44×10⁷ m³/5 yr in 2000-2020. During the same period, the river water consumption between the Daxihaizi Reservoir and the Alagan section was 2.79×10⁸ m³/yr. Vegetation water consumption accounted for 68.23% of the river water consumption, and both increased with the water conveyance process. ③ The natural vegetation area increased at a rate of 12.37%/5 yr, and water consumption increased at a rate of 12.82%/5 yr. The vegetation area matched the growth rate of water consumption. The restoration of natural vegetation responded positively to water conveyance, and continued water conveyance will still promote the rapid restoration of vegetation.[Conclusion] By 2020, after 21 years of ecological water conveyance, the natural vegetation in the lower reaches of the Tarim River has been significantly improved. However, due to the solidification of existing water conveyance methods, water losses by evaporation from the water surface and shallow buried depth areas have increased. The temporal and spatial distribution of ecological water consumption is still feasible and should be optimized.