Abstract: For the tower structure of an offshore articulated wind turbine,a multi-objective optimization mathematical model is established by analyzing its mechanical properties and coupling relationship between structures under complex environment.Meanwhile,a non-dominated sorting multi-objective genetic algorithm based on spatial reference points is used to optimize dimensions of tower parameters. Then some fuzzy evaluation indexes are proposed and the sensitivity analysis of optimization parameters is also carried out according to Pareto solutions to obtain the optimal solution scheme. The calculation results show that the maximum equivalent stress,top displacement and buckling safety factor of the tower meet the safety operation permit under the rated wind speed sea state.And compared with the NREL 5 MW prototype tower structure,the overall mass decreases by 13.41%,the equivalent stress decreases by 13.79%. However,the maximum displacement of tower top increases by 3.36%,it is still far less than the safety offset which meets the design requirements. The buckling safety factor of the tower is 0.05,which is far less than 1 and meets the buckling stability requirements. At the same time,the natural frequency of tower bending vibration is not in the range of 1P and 3P of blade wind wheel rotation,which can avoid the occurrence of structural resonance. The tower structure design based on multi-objective optimization has been significantly improved both in structural performance and production cost.