Abstract: The optimal Bang-Bang control problem of low-thrust space trajectories is introduced. Two classes of continuation methods are described:the first solves the energy-optimal solution,and subsequently employs the energy-fuel homotopy to obtain the optimal Bang-Bang control;the second introduces a switching principle,and obtains the optimal Bang-Bang control through parameter continuation starting from a two-impulse solution. The two continuation methods are compared,and the advantages and characteristics of the two methods are discussed. The prospects of the continuation methods applying to more complicated low-thrust trajectory designs are proposed. The concept of artificial intelligence trajectory optimization is presented,which contains three aspects:initial solution,continuation,and patching.