The wind field retrieved from a Doppler radar largely contributes to improvement of initial data in mesoscale numerical weather prediction model as well as better understanding of structure of local severe weather systems due to its high temporal and spatial resolution. In this study, a technique for retrieving horizontal wind field from the KMA (Korea Meteorological Administration) radar data was developed by assimilating the gridded and objectively-analyzed radar data into the first guess values obtained from numerical weather prediction model. The assimilation method used is a simple successive correction method introduced in KLAPS (Korea Local Analysis and Prediction System). Also, a comparison with radiosonde-observed wind data was made to assess the accuracy of the retrieved wind data. Furthermore, the folding problem of radial velocity and its quality control was examined prior to the development of retrieval technique of wind data from a Doppler radar.