The Canadian prairie region is known for its complex cold-region hydrology and flat terrain with many depressions, which intercept runoff through the fill and spill mechanism. This process contributes to non-linear hysteretic relationships between water storage and the contributing fractions of prairie basins. Current hydrological models either ignore or poorly represent such a complex phenomenon, and therefore limited success has been achieved so far in simulating the streamflow of this area. In this study, a new Hysteretic Depressional Storage (HDS) algorithm is proposed to simulate the fill and spill mechanism and the associated variable contributing area and is implemented in the HYPE modelling framework. The modified HYPE model (HYPE-HDS) and the original version of the model with the traditional lake module (HYPE-ilake) are tested on two depression-dominated prairie basins. The HYPE-HDS model shows an improved streamflow simulation compared to the original HYPE-ilake model. Most importantly, HYPE-HDS replicates the complex hysteretic relationships between depressional storage and basin contributing fractions, unlike HYPE-ilake model that cannot produce such relationships. The HDS implementation approach proposed in this study is designed for use in any hydrological or land surface modelling system, to improve its ability to simulate the complex prairie hydrology and streamflow. This approach should also improve the estimates of surface water storage and the resulting evaporative fluxes, which should improve model fidelity and water budget estimates in this complex environment.