Soil moisture is a key link between the land surface and atmosphere, influencing water balance and major hydrological processes. This study examined soil moisture patterns and their relationship with soil properties, precipitation, and evaporation in the Yanting agroforestry watershed. Three cropland sites with hilltop, middle slope, and lower slope were selected to measure the soil moisture contents (SMCs) at various soil depths from 2015 to 2017 for adjustment and validation of HYDRUS-1D. The model was used to simulate changes in SMCs with the interval of five years from 2020 to 2050 under two climate scenarios (RCP 4.5 and RCP 8.5). The dynamic changes in soil moisture were captured at different slope positions during calibration (R-2 >= 0.94; RMSE <= 0.021 cm(3)/cm(3)) and validation (R-2 >= 0.90; RMSE <= 0.009 cm(3)/cm(3)). Soil moisture variability followed the pattern: lower slope > middle slope > hilltop. The projected SMCs of hilltop and middle slope cropland reported a surplus in 2030. Between 2020 and 2050, croplands situated on the hilltop, middle slope, and lower slope experienced 46, 140, and 217 days per year when average SMCs exceeded the field capacity, respectively. The hilltop cropland found the highest optimal number of days per year, suggesting that in future, hilltop cropland will be more suitable for crop production. The results also showed that SMCs vary with soil depth (20 to 60 cm) and exhibited higher spectral and temporal variability at the surface compared to the deeper soil layer. These conclusions could help irrigation water managers and hydrologists in the development of successful and sustainable agroforestry systems in watersheds to protect land and water resources.