Rainfall has two significant effects on the sea surface, including salinity decreasing and surface becoming rougher,which have further influence on L-band sea surface emissivity. Investigations using the Aquarius and TRMM 3B42 matchup dataset indicate that the retrieved sea surface salinity(SSS) is underestimated by the present Aquarius algorithm compared to numerical model outputs, especially in cases of a high rain rate. For example, the bias between satellite-observed SSS and numerical model SSS is approximately 2 when the rain rate is 25 mm/h. The bias can be eliminated by accounting for rain-induced roughness, which is usually modeled by rain-generated ring-wave spectrum. The rain spectrum will be input into the Small Slope Approximation(SSA) model for the simulation of sea surface emissivity influenced by rain. The comparison with theoretical model indicated that the empirical model of rain spectrumis more suitable to be used in the simulation. Further, the coefficients of the rain spectrum are modified by fitting the simulations with the observations of the 2–year Aquarius and TRMM matchup dataset. The calculations confirm that the sea surface emissivity increases with the wind speed and rain rate. The increase induced by the rain rate is rapid in the case of low rain rate and low wind speed. Finally, a modified model of sea surface emissivity including the rain spectrum is proposed and validated by using the matchup dataset in May 2014. Compared with observations, the bias of the rain-induced sea surface emissivity simulated by the modified modelis approximately 1e–4, and the RMSE is slightly larger than 1e–3. With using more matchup data, thebias between model retrieved sea surface salinities and observationsmay be further corrected,and the RMSE may be reduced to less than 1 in the cases of low rain rate and low wind speed. Investigations using the Aquarius and TRMM 3B42 matchup dataset indicate that the sea surface salinity (SSS) is retrieved underestimated by the present Aquarius algorithm compared to numerical model outputs, especially in cases of a high rain rate. For example, the bias between satellite-observed SSS and numerical model SSS is approximately 2 when the rain rate is 25 mm / h. The bias can be eliminated by accounting for rain-induced roughness, which is usually modeled by rain-generated ring-wave spectrum. The rain spectrum will be input into the Small Slope Approximation (SSA) model for the simulation of sea surface emissivity influenced by rain. The comparison with theoretical model indicates that the empirical model of rain spectrum is more suitable to be used in the simulation. Further, the coefficients of the ra in spectrum are modified by fitting the simulations with the observations of the 2-year Aquarius and TRMM matchup dataset. The calculations confirm that the sea surface emissivity increases with the wind speed and rain rate. The increase induced by the rain rate is rapid in the Finally, a modified model of sea surface emissivity including the rain spectrum is proposed and validated by using the matchup dataset in May 2014. Compared with observations, the bias of the rain-induced sea surface emissivity simulated by the modified modelis approximately 1e-4, and the RMSE is slightly larger than 1e-3. With using more matchup data, thebias between model retrieved sea surface salinities and observationsmay be further corrected, and the RMSE may be reduced to less than 1 in the cases of low rain rate and low wind speed.