Abstract: According to the high precision demand of estimating TOA in Galileo Search And Rescue (SAR) system, and considering the uncertainty of message bit width and the asymmetry degree of modulation in actual COSPAS-SARSAT signal, a new TOA estimation algorithm based on Multiple dimensions Joint Maximum Likelihood Estimation (MJMLE) algorithm is proposed. The COSPAS-SARSAT signal model containing the asymmetry degree of modulation is introduced first. Then the principle of the algorithm is derived and the concrete realization is given. Monte Carlo simulation results show that when Carrier to Noise Ratio (CNR) equals to the threshold of 34.8 dBHz and the asymmetry degree of modulation is the extreme condition with 5% and -5%, root-mean-square errors (rmse) of TOA estimation of this algorithm is less than 13.5 s, which can satisfy the system requirement of 15 s and better than that of other estimation algorithms.