Ocean waves are essential elements across the air‐sea interface, regulating momentum and energy transfer. The mixture of wind sea and ocean swell coupled with surface winds results in diverse sea state conditions that modify the local air‐sea interaction. Previous classifications of wind waves and swells are mostly binary that are insufficient to represent the complexity of sea states. In this study, we utilize wind and wave measurements from the China‐France Oceanography Satellite (CFOSAT) to construct an observational wind‐wave ensemble. Four key parameters: wind speed, significant wave height, inverse wave age, and spectral width are selected out of six variables based on their correlations. Employing the unsupervised learning of k‐means clustering, global sea states are categorized into six distinct classes. These classes, characterized by unique centroids and separated in the feature space, represent specific wind regimes and degrees of wave development. Global occurrence highlights that each sea state is region‐specific, bridging the spatial gap of swell and wind sea dominated areas, respectively. This new grouping scheme complements the traditional wind sea and/or swell classification by resolving the diversity of wave regimes. The six‐class classification enables us to identify transitional states and hybrid conditions that may have been overlooked in the binary classification scheme, which shall help investigate the impact of ocean waves on the air‐sea interaction under varying sea states.