The extraction of full wind vectors from data obtained by single-site (monostatic) high-frequency ground-wave radar (HFGWR) is an ongoing challenge because of the inherent directional ambiguities. Here, a new algorithm for resolving the ambiguity of wind direction from monostatic data is presented. The true wind direction is determined by minimizing the sum of the difference among three wind directions derived from three different radar look angles. The wind directions estimated by applying the algorithm to data obtained from the OSMAR2000 HFGWR situated at the Eastern China Sea are compared with values obtained from ship-borne instrumentation. The mean difference between the ground-truthed values and those obtained from the radar data is approximately 20°. The distance limit for wind direction sensing using the OSMAR2000 is about 200 km, which is the range for which signal-to-noise ratios typically exceed about 23 dB in the relevant first-order portions of the backscatter spectra.