Gravel mulching plays a vital role in modifying the hydrological cycle in arid and semi-arid areas. Yet, the mechanisms underlying long-term mulching effects on soil evaporation remain poorly understood. To investigate the hydrological effects of mixed gravel-soil mulching (MGSM), we conducted a controlled 39-d soil evaporation experiment (from 22 July to 30 August 2021) using micro-lysimeters at the field experimental site of Ningxia University, China. The soil evaporation rate (E), cumulative soil evaporation (E-c), soil water content (SWC), mulch resistance (r(m)), and micro-meteorological variables were assessed for six mulch treatments, each containing a different proportion of gravel by volume: 100.00% (M1), 80.00% (M2), 60.00% (M3), 40.00% (M4), 20.00% (M5), and 0.00% (M6). The treatments (M2-M6) showed a prolonged soil moisture depletion phase and greater E-c (28.71%-83.31%) relative to the gravel-only treatment (M1) (P<0.050); these effects were primarily attributed to reduced r(m). As compared to E-c, the SWC showed an inverse response, decreasing as Cg decreased. A robust exponential relationship was observed between E and r(m) (P<0.001). Evaporation suppression mediated by r(m) was particularly pronounced during the residual evaporation stage (>312 h post-wetting), with the strongest effect occurring in M3, where the mean r(m) doubled. The SWC, mulch properties, and micro-meteorological parameters (i.e., air relative humidity and surface net radiation flux) were the most important predictors of r(m) in the mulch treatments. Together, these results suggested that MGSM unexpectedly exacerbated surface soil moisture loss by reducing r(m). To mitigate this effect, an optimized mixed gravel-soil mulch, containing 60.00% gravel by volume, might be used; this mixture balances evaporation control with hydrological sustainability and represents a practical strategy for dryland management, offering a compromise between short-term water retention and sustained soil moisture regulation.