We introduce a new proxy measure for VS30 (time‐averaged shear‐wave velocity in the upper 30 m) for central and eastern North America (CENA). The new proxy is the site fundamental frequency (fpeak), measured from the horizontal‐to‐vertical (H/V) spectral ratios of recorded ground motion (or ambient noise). In this study, H/V spectral ratios are obtained from 5% damped pseudospectral acceleration (PSA) from seismograph stations in CENA using the Next Generation Attenuation‐East (NGA‐East) database. We correlate the measured VS30 values at recording stations with the corresponding fpeak values to obtain a predictive relationship for VS30. The uncertainty of the VS30 estimate using the fpeak‐based model is small (0.14log10 units) in comparison to that for the proxy‐based methods (e.g., topographic slope and surface geology proxies) used in the NGA‐East database (0.25log10 units). However, values of fpeak can be obtained for only about 23% of the NGA‐East recording stations due to data and site‐type limitations. Reducing the error in VS30 estimates can potentially reduce the variability of ground‐motion prediction equations (GMPEs). To test this hypothesis, we use the fpeak‐based VS30 estimates and select one of the proponent NGA‐East GMPE models. For a selected database, we are able to reduce the GMPE variability (σ) by 3% on average, for PSA at 1–10 Hz, just using the fpeak‐based proxy to estimate VS30. Greater variability reductions could be achieved by replacing VS30 with improved site characterization parameters, including fpeak, in the GMPEs.