The prediction of streamflow helps to identify the disasters and sources of water resources. Different ways to predict the streamflow, among the conceptual models have been gained more popularity due to explanation of processes and visualization in water resources systems. Any model may not obtain the acceptable performance at initial setup and it has to go through calibration or optimization (either manual or auto-calibration). Moreover, the calibration procedure is more concern of computational time for complex conceptual models like Soil and Water Assessment Tool (SWAT). Where, meta models are the alternative approach to restrict the computationally intensive optimization problems because it is cheaper models to enhance the performance and shows the relationship between input-output response. Our results showed that 1) meta models mimics the original simulation models with effective and efficient outputs and 2) it verified and satisfied the performance of SWAT model with less computational time. This study helps to planning and designing of hydrological models with effective computational time. © 2016 Author.