Parametric spectrum models describing a frequency spectrum of ocean surface waves (e.g., Pierson-Moskowitz and JONSWAP (Joint North Sea Wave Project) spectrum) as a function of the significant wave height and peak period (Tp) have been used to estimate the energy period (Te) using a relation, Te=CTp, when the frequency wave spectrum is not available. This approach, however, introduces uncertainties in Te estimation for irregular waves. In the present study, a regression model for defining the coefficient (C) is developed to minimize the uncertainty in Te estimation using buoy wave measurements and WaveWatch III® hindcast for coastal waters of the United States. The result shows that the coefficients (C) of wind sea, swell, and total sea vary geographically and range from 0.84 to 0.90, 0.92 to 0.98, and 0.79 to 0.85. This approach derives consistent coefficients with different periods of record and results in better agreements in Te estimation compared to those of the approach relying on parametric spectrum models. Frequency distributions of dominant wave systems reaching different coastal regions, e.g., Pacific northwest swell in the Pacific Ocean, nor'easter swell and trade wind swell in the Atlantic Ocean, explain the observed spatial trends in the coefficient.