<p>In the semi-arid region of Morocco, water resources are limited and under severe strain. In this environment, watersheds are heterogeneous, have nonlinear hydrological responses, and are characterized by data scarcity. Considering these particularities, using an appropriate modeling approach that considers the unique characteristics of semi-arid watersheds with limited data is crucial for effective and sustainable water management. The physically based SWAT model combined with SWAT-CUP and the low-parameterized GR4J lumped model were used to that end. These models simulated daily observed streamflow and hydrological signature indices to investigate the utility of lumped and spatially distributed modeling approaches in reproducing the hydrological behavior of the Oued Issen watershed. Our findings demonstrated that both modeling approaches performed well (NSE &gt; 0.50 and R² &gt; 0.60). Furthermore, the comparison results indicated that the spatially distributed modeling approach outperformed the lumped modeling approach. The latter was less consistent due to its inability to account for spatial variability in precipitation and the hydrological characteristics of the Oued Issen watershed. Compared to the GR4J model, the SWAT model illustrated its superiority by more accurately reproducing the watershed’s hydrological response, specifically in terms of streamflow magnitude. Using the SWAT-CUP program reduced SWAT model uncertainties and avoided the risk of overparameterization. Using the semi-distributed SWAT model with SWAT-CUP proved to be the most practical approach for studying semi-arid Mediterranean watersheds, balancing process complexity with data scarcity. Overall, modelers can rely on this approach for better water resource management in data-scarce, semi-arid regions.</p>