Abstract: The wind speed distribution in offshore wind farms is an important factor for evaluating the generated capacity. The distribution of short-term wind speed in offshore wind farms always has multi-peak characteristics,so the mixed Weibull distribution is required. Due to the variety of parameters and the difficulty in making the approximation of this distribution,a mixed Weibull distribution approximation method,which takes the minimum variance as the optimization objective and the value range of each parameter as the constraint condition,is proposed in this paper. Further,a modified particle swarm optimization algorithm,DAIWtanh,is used to optimize the parameters in the proposed method,which can enhance the approximation accuracy of multi-peak wind speed distribution. The numerical results show that the method is able to achieve higher approximation accuracy and higher computation speed,and is not easily trapped in local optimization,leading to a better approximation effect.