The SCE-UA global optimization method was used for the Xinanjiang model parameter calibration on the Yuetan basin.The results reveal that the parameter optimization results are stable when using the ideal calibration data.However, for the observed calibration data,SCE-UA method can't get the only and stable parameter set for sure;In the Xinanjiang watershed model for daily simulation,the Water Quantity Balance Error and Deterministic Coefficient are two best objective functions and in the Xinanjiang watershed model for hourly simulation,the Absolute Logarithm Error is the best objective function;In order to get the stability pa-rameter set,more than 12 years' actual data is needed;then the objective optimization theory which advanced by ZHAO Renjun has been introduced and combined with the SCE-UA method, and the results reveal that the parameter optimization results are more sta-ble than the way that didn't use this theory.