Hurricane Hugo struck Charleston, South Carolina, on 22 September 1989 as the most intense hurricane to affect the United States since Camille in 1969. The northeastern eyewall, which contained the maximum winds measured by reconnaissance aircraft shortly before landfall, moved inland over a relatively unpopulated area and there were few fatalities. However, no observations were available to document the surface wind distribution in this part of the storm as it continued inland. To improve specification of surface winds in Hugo, empirically adjusted aircraft winds were combined with coastal, offshore, and inland surface observations and were input to the Ooyama objective analysis algorithm. The wind analysis at landfall was then compared with subsequent analyses at 3 and 6 h after landfall. Reconstruction of the surface wind field at landfall suggests that the maximum (∼13 min mean) surface wind at the coast was 50 m s−1 in the Bulls Bay region, ∼40 km northeast of Charleston. Surface roughness over land caused wind speeds to drop off rapidly just inland of the coast to only 50% of values measured by reconnaissance aircraft at the same location relative to the storm over water. Despite relatively rapid increases in the central sea-level pressure and decreases in the mean circulation as Hugo progressed inland, hurricane-force wind gusts extended Hugo's damage pattern well past Charlotte, North Carolina, ∼330 km inland. Accurate determination of surface wind distribution in land-falling hurricanes is dependent upon the spatial density and quality of surface wind measurements and techniques to adjust reconnaissance flight-level winds to the surface. Improvements should allow forecasters to prepare more-accurate warnings and advisories and allow more-thorough documentation of poststorm effects. Empirical adjustments to reconnaissance aircraft measurements may replace surface data voids if the vertical profile of the horizontal wind is known. Expanded use of the airborne stepped-frequency microwave radiometer for remote sensing of ocean surface winds could fill data voids without relying upon empirical methods or models. A larger network of offshore, coastal, and inland surface platforms at standard (10-m) elevations with improved sampling strategies is envisioned for better resolution of hurricane wind fields. A rapid-response automatic station network, deployed at prearranged coastal locations by local universities with meteorology and/or wind engineering programs, could further supplement the fixed platform network and avoid the logistical problems posed by sending outside teams into threatened areas.