Since 1997 typhoon season, BATS (Barotropic Adaptive-grid Typhoon Simulation) model along with GFDK (GFDL model in KMA) and KTM (Korea Typhoon Model) and GDAPS (Global Data Assimilation and Prediction System) has been put into operation to produce the typhoon forecast for KMA (Korea Meteorological Administration). KTM is a baroclinic regional spectral model and GDAPS is a global spectral model. Although being a barotropic model, BATS has been quite successful in forecasting typhoon track. Extensive effort has been put to improve the performance of the BATS model (Kwon and Shin 1996, Kwon and Park 1997). Up until now, the steering current has been taken directly from the reported motion in 'tropical cyclone advisory' of RSMC Tokyo, when bogussing the vortex. The motion vector here is generally from the past six hour motion and converted into every sixteenth azimuthal resolution. However an examination shows that this motion vector has error of about one azimuthal resolution comparing to the actual motion. Therefore, another way of defining the steering current is proposed. If the previous steering is based on 6 hour PERsistence vector, 6 hour CLIPER forecast vector would give more accurate diagnosis of the motion vector. In addition, the model is integrated for 12 hour to see whether the initially given vortex moves according to the prescribed steering current. Then, the steering current is modified to force the vortex to move to the prescribed steering current. The results are promising in the sense that the forecast track errors are reduced for almost every individual storm case