The design optimization of hypersonic vehicle back sketch lines was investigated. A parametric geometry model was presented based on free form deform (FFD) method and Hicks-Henne function

which was capable of keeping the back surface smooth and continuous during deforming. In order to save time cost

unstructured grid space marching method was adopted to solve the Euler equations. The accuracy and efficiency of this method were analyzed by DOE methods through hypersonic cone cases. There was a good conformity between the un-structured grid space marching and time marching results of the hypersonic vehicle flow. Indicated by the comparisons

the unstructured grid space marching method had the capacity to increase the computational efficiency over 4.5 times. A response surface model was built through optimize latin hypercube sampling. The Pareto effects of design variables were analyzed. The multi-objective optimization algorithm NSGA-Ⅱ (non-dominated sorting genetic algorithm-Ⅱ) was adopted to optimize the drag coefficient and lift-to-drag ratio under the constraint of the vehicle volume change within 20%. The optimization history showed a negative correlation between the drag coefficient and lift-to-drag ratio. The drag coefficient of the optimal configuration decreased by 7.1% and the lift-to-drag ratio increased by 16.9%.