Most of the wave information systems are based on the spectral representation of the observed wave data. Spectral analysis describes the wave energy distribution as a function of frequency for a given wave data set. However, spectral analysis may not provide the joint statistical distribution of wave heights and periods. An alternative spectrum estimation by means of the joint probability density function of wave heights and periods is presented. The wave height joint probability spectra are defined as the second moment of wave height of joint probability in terms of the frequency domain. Good agreement between the joint probability density spectra and fast Fourier transform and autoregressive spectra was obtained.