Abstract: Estimating the importance measure of a structural system by global sensitivity analysis methods can figure out the importance ranking of the input random variables that affect the uncertainty of the output response and the risk level, which provides a basis for the subsequent research on reliability evaluation, on fault diagnosis and system design, on prediction and optimization. A variety of global sensitivity methods has been developed. However, the global sensitivity analysis on high dimensional and complex structures such as wind turbine blade is still a problem to be solved. Three possible calculation forms of space partition global sensitivity analysis methods and the optimal partition scheme are thusly studied. The optimal calculation form and partition scheme, being able to utilize the information of samples and to reduce computational burden as fully as possible, are presented via the analysis of the standard example and by reasoning of the errors. The method is applied to the global sensitivity analysis of wind turbine blade under the ultimate load condition. The method has demonstrated its significance in improving the efficiency on future design of the wind turbine blade, on saving design cost and improving product quality.