Water resource management is critical in many arid environments. The understanding and modelling of hydrological systems shed light on important factors affecting scarce water resources. In this study, a semi distributed hydrological model capable of simulating water balance in large geographical catchments and sub basin was used for runoff estimation in the Okavango Omatako catchment in Namibia. The model was configured for a thirty–one year period from 1985 to 2015 as per the availability of data for the study area. Subsequently, calibration and validation processes followed for the period 1990-2003 (calibration) and 2004-2008 (validation) using the sequential uncertainty fitting 2 (SUFI 2) algorithm. For the evaluation of the simulation of the Okavango Omatako catchment, two methods were used: i. model prediction uncertainty and ii. model performance indicators. Prediction uncertainty was used to quantify the goodness of fit between observed and simulated result of model calibration, which is measured by P factor and R factor. The P-factor achieved 0.77 during calibration and 0.68 for the validation. The value for calibration was adequate while the validation value was around the recommended value of 0.7. On the other hand, the R factor attained 1.31 in the calibration and 1.82 during validation. The calibration result was within the acceptable range while the validation was slightly on the upper side. The following indicators were used to evaluate the model performance through calibration and validation results respectively; Nash Sutcliffe Efficiency (NSE) with 0.82 and 0.80, Coefficient of determination (R^2) with 0.84 and 0.89, Percent bias (PBIAS) achieving -20≤PBIAS≤-1.1 and residual variation (RSR) performing 0.42 and 0.44. All performance indices achieved very good ratings apart from PBIAS validation which rated as satisfactory. It is therefore recommended to use SWAT for semi-arid streamflow simulations as it demonstrated reasonable results in modeling high and low flows.