Based on the effective medium approximation theory of composites, the empirical model proposed by Pandey and Kakar is remedied to investigate the microwave emissivity of sea surface under wave breaking driven by strong wind. In the improved model, the effects of seawater bubbles, droplets and difference in temperature of air and sea interface (DTAS) on the emissivity of sea surface covered by whitecaps are discussed. The model results indicate that the effective emissivity of sea surface increases with DTAS increasing, and the impacts of bubble structures and thickness of whitecaps layer on the emissivity are included in the model by introducing the effective dielectric constant of whitecaps layer. Moreover, a good agreement is obtained by comparing the model results with the Rose's experimental data.wave breaking, whitecaps, emissivity, effective medium approximation, radiant intensity Because of the importance of wave breaking in theoretic researches and engineering applications, it had attracted much attention, such as Munk [1] observed the whitecaps coverage under wave breaking early. Blanchard [2] gave the empirical relationship of whitecaps coverage and the 10 m high wind speed above the sea level. With the analysis of measurement data of the East China Sea and Pacific Ocean, Toba and Chean [3] also gave the empirical relationship of whitecaps coverage and wind speed. Ross and Cardone [4] obtained the whitecaps coverage of different wind speed from space photography of the sea surface state. Xu [5] improved the measurement method of wind-wave breaking, which included the measurement of broken-wave intensity. These relationships between whitecaps coverage and the 10 m high wind speed were directly obtained from measurement of sea surface. However, it is very difficult to measure other important parameters of sea-air interface under wave breaking. Recently, with the development of ocean remote sensing, the modern remote sensing technology has been used to inverse the physical parameters of the sea-air interface [6][7][8] . In order to estimate the effects of the sea-air interface layer on remote sensing measurement, many remote sensing models have been applied to the sea-air interface [9][10][11][12][13][14][15][16][17][18][19][20][21][22] . For example, the sea surface backscattering models were studied. Wu and Fung [9] proposed the two-scale non-coherent emissivity and backscattering models. Jin et al. [10] discussed the backscattering model of rough sea surface with foams. McDaniel [12] derived the backscattering model of non-Gaussian sea surface. In addition, the sea surface microwave emissivity models were proposed by many authors. For example, with least-square method, Stogryn [14] studied whitecaps emissivity model, which included the effects of incident angles and frequencies. Pandey and Kakar [15] had gotten the semi-empirical emissivity model of rough sea surface. Smith [16] gave the brightness temperature model of foam-covered sea surface. Huang and Jin [17] derived the whitecaps emis-