Plants on the land surface play a vital role in the hydrological water cycle as they transport soil water to the atmosphere through transpiration. Root water uptake (RWU) is considered a crucial step in this process as it is the first stage of transpiration, directly determining the actual transpiration (Ta) of plants. However, accurately measuring RWU or Ta in situ poses significant challenges. Here, we establish an overall approach of combining mathematical models and machine learning algorithms to obtain high-precision (500 m×500 m) regional-scale daily Ta maps for various future climate patterns. The Hydrus-1D and AquaCrop models were employed to calculate the total RWU fluxes across the entire root zone, aiming to achieve Ta at a point scale. A machine learning model was developed using the CatBoost algorithm and environmental covariates extracted from the Google Earth Engine (GEE) platform to upscale these point-scale Ta to the regional scale. Furthermore, a total of 22 CMIP6 Earth System Models (ESMs) were evaluated, and among them, ACCESS-CM2 and ACCESS-ESM1–5 were selected for simulating future climate scenarios. Based on the established machine learning model and selected ESMs, regional-scale Ta maps were generated from 2020 to 2100 for the SSP245 and SSP585 (Shared Socioeconomic Pathways) scenarios. The results indicate that near-surface specific humidity, mean near-surface air temperature, latitude, and surface downwelling shortwave radiation are the critical factors influencing regional-scale Ta. As greenhouse gas emissions intensify and temperatures rise, regional-scale Ta is enhanced, leading to an accelerated transfer of soil water to atmospheric water. Under the SSP245 scenario, Ta increases on average by 0.55–1.16 % every 20 years, with its incremental value ranging from 7.14∙10−4 to 8.65∙10−4 cm day−1, while under the SSP585 scenario, Ta increases more significantly, achieving an average increase of 0.64–1.81 % every 20 years, with its incremental value ranging from 1.595∙10−3 to 2.821∙10−3 cm day−1. This study provides a robust integrated approach to assess the future regional-scale Ta providing valuable insights into the underlying water cycle mechanisms and regional water requirements for future climate scenarios.