The internal adjustment process of a hydrological model followed by an unusual initial condition is known as the model spin-up. And the time required for a complete adjustment is termed as the model spin-up time. Model results for the duration of this spin-up progression are greatly impacted by the initial conditions, and often impractical or erroneous. The speed of this adjustment process is affected by the characteristics of the input data sets and their persistence. This study discusses the variability and seasonality of hydrological model spin-up time against the aridity of the river basin using multi-year climatologies for 18 river basins distributed relatively snow-free regions of the USA. The Xinanjiang model was run with each of all available year input data sets with two extreme initial conditions (saturated and unsaturated) and thereafter detected the model equilibrium state based on the Mahalanobis distance between the soil moisture states of two model runs. The seasonality of model spin-up was investigated by conducting multiple simulations that start from different time of a year. The basin average soil moisture memory (SMM) timescale (Rahman et al., 2015) and basin aridity index was estimated and thereafter investigated their relationship with the average model spin-up time. <br><br> Analysis suggests that the spin-up time highly varies with the simulation starting time and the dryness of the river basin. Overall, in all basins, model achieves the equilibrium state quickly while the simulation starts in late autumn (October–November). On the other hand, model equilibrates slowly while simulation starts in spring (March–May). Wet basin shows stronger variability of the model spin-up time (mean range 154 days) throughout the year as compared with that of dry basins (mean range 78 days). The mean spin-up time is shorter for wet basins (154 days) and longer for dry basins (233 days). The spin-up times are 3–7 times longer than the SMM timescale. The basin-wise mean spin-up time shows linear and exponential relationship with the SMM timescale and the basin aridity index respectively. The relationship offers predictability of model spin-up time from widely available potential evaporation and precipitation data sets.