The SWAT model (Soil and Water Assessment Tool) has been widely used. It is a semi-distributed hydrological model built to simulate the effects of land use management and climatic factors on water source, sediment and organic matter content in river basin system with soil types, with different land use conditions for a period or forecast for a long time. Besides its outstanding advantages, the SWAT model also requires a large number of input parameters, the process of parameterizing, calibrating and verifying the model with manual methods is facing many difficulties in the implementation process. This is an extremely important step to evaluate the accuracy of simulation model results, however, with a large number of parameters and complexity across sub-basins, changing and detection each parameter to determine the sensitivity in the process of calibration and validation really takes a lot of time and affects the quality of the model. Therefore, in this study, it is necessary to focus on detailed research on the method of auto-calibration and validation the model to reduce time, volume and improve the quality of the model to determine the highest sensitivity parameters and influence in the model The SWAT model result was calibrated and validated by comperison of observed data against simulated of monthly streamflow where parameters were adjusted based on the sensitivity analysis in auto-calibration methods. The research results provide a useful method with powerful tools, saving time and effort to increase the quality, consistency and accuracy of the model.