Sea surface wind is a key parameter in the studies of many oceanic applications, i.e. meteorological forecasting, oil slick observation, ship detection, wind turbine installation recently, etc. Among available wind sources, i.e. in situ measurement, numeric weather models, etc., the retrieval of wind fields from Synthetic Aperture Radar (SAR) data is selected in this study, since it may give winds at high spatial resolutions. For this approach, wind directions are first retrieved from SAR images by the Local Gradient method. They are then used in the Geophysical Model Functions (GMFs) to estimate wind speed. The GMFs are used to describe the dependency of radar scattering from sea surface on surface wind speed and the geometry of observations. Therefore, by knowing radar scattering and geometric parameters, it is possible to invert GMFs to estimate wind speed. There are two principal approaches to model GMFs: based on empirical (EP) descriptions (so called EP GMFs), and based on electromagnetic (EM) calculations (so called EM GMFs). For an overview, wind speeds are estimated using both EP and EM models. They are then compared with in situ measurements to evaluate. Based on the comparisons, the advantages and inconveniences of the GMFs are analyzed.