Digital elevation model (DEM) resolution influences hydrological simulation. However, its influence when modelling hydrological connectivity (HC) in woody crops remains to be seen. We assessed surface topography, microtopography and HC in an agricultural sub‐catchment (27.4 ha) using six photogrammetry‐derived DEMs with 0.03, 0.05, 0.1, 0.2, 0.5 and 1 m cell sizes. Land uses included vineyards, olive groves, cereal fields, and forestry patches. We ran an updated version of Borselli's HC index (IC) using the D‐Infinity approach. We assessed HC in woody crops at high spatial resolution for the first time. After analysing the differences in the contributing area, the flow width, the soil roughness, the convergence index and the IC (normalised and non‐normalised) at different scales (hillslope, land uses and compartments, ephemeral gullies, depositional areas and the sub‐catchment outlet) and accounting for the field vertical components, we propose an optimum DEM resolution (0.2 m) to improve modelling of structural HC in woody crops. The modelled hydrological features at this threshold resolution matched well with the geomorphic features associated with the short‐ and medium‐term patterns of soil redistribution. Higher DEM resolutions, especially at 0.03 and 0.05 m, introduced bias in the input data and the IC computations. Finally, we observed good agreement between the outputs at the lowest resolution, 1 × 1 m, and the long‐term soil redistribution patterns (functional connectivity). Copyright © 2017 John Wiley & Sons, Ltd. For the first time we demonstrated the existence of a threshold DEM resolution at 0.2 m to improve model predictions of structural hydrological connectivity (Borselli's index – updated version) and surface roughness in woody crops (vineyards and olive groves). Higher DEM resolutions, especially at 0.03 and 0.05 m, introduced bias in the input data. Different connectivity patterns and values were observed in the vineyards' rows and inter‐rows. Long‐term functional connectivity was better modelled at low DEM resolutions (0.5 and 1 m).