Understanding spatio-temporal variability of soil water content (SWC) at different scales is of great importance for designing monitoring networks and assimilating observed data in hydrological modeling. The objective of the study was to examine spatio-temporal variability of SWC for different soil layers at different extent scales within an agricultural field. Spatio-temporal variability of SWC and the relationship between variability and mean were investigated at three extent scales for different soil layers of 0–100 cm in different years within an intensively irrigated vineyard. Nested sampling grids were used, and at each scale, there was a total of 16 points distributed regularly at the grids. The distance between any two adjacent sampling points was 25-m, 50-m, and 75-m for the three scales respectively. Overall, there was no consistent pattern regarding the change of mean, standard deviation (SD), and coefficient of variation (CV) with scales during the three years. For the surface soil layer, mean SWC was similar between scales, while SD was closely related to soil clay content at the sampling locations of each scale. For subsurface soils, the overall magnitude of variability of SWC was greater, whereas the difference in the variability between scales was lower than that of the soil texture of the corresponding soil layer. The proportional decrease of CV with increasing mean was found to be largely consistent between different years for all soil layers excluding 20–40 cm, but variable between scales and soil layers. For subsurface soil layers, the CV vs. mean relationship differed between relatively wet and dry years, indicating that factors controlling SWC dynamics for the layers below the surface layer were different depending on climatic conditions. Variance partitioning analysis showed that spatial variance accounted for at least 57 % of total variance for all cases, and the temporal variance at different scales fluctuated within a narrow range for subsurface soils for the three years. The results could benefit soil water content monitoring network design, calibration of soil moisture related parameters and assimilation of observed soil moisture data into field soil water modeling.