Uncertainty is inherent in modeling studies. However, the quantification of uncertainties associated with a model is a challenging task. Furthermore, snowmelt estimation is a crucial part of the Soil and Water Assessment Tool (SWAT) model in watersheds where spring runoff is strongly affected by melting snow. The SWAT model for the snow dependent Kunhar basin in Himalayan watershed was calibrated (2001 to 2005) and validated (2006 to 2009) using Sequential Uncertainty Fitting Algorithm (SUFI-2). For the model uncertainty two indices P-factor and R-factor along with frequently used objective functions R2, NSE, PBIAS, were taken into consideration. For calibration, multisite daily and monthly simulation results of SUFI-2 revealed that percentage of data enveloped by 95% confidence interval was 85% (monthly) to 87% (daily) at upstream calibration point and 63% (monthly) to 73% (daily) data at the downstream calibration point. Model validation by the usage of elevation bands indicated better model performance, enveloping 15% to 20% more observed data than the validation without elevation bands together with the other statistical standards. Equifinality in the model parameters was observed, and it was discovered that the model uncertainty lie inside the model parameters. It is recommended that critical model parameters correspondence with the watershed characteristics should be checked. The calibrated version of the model could be further used for the analysis and impacts of climate and land use changes on stream flows, water quality and sediment yield.