Based on stochastic time and fuzzy loss in war, the model of optimal path is constructed by dependent-chance programming and a genetic algorithm is designed. For solving the problems of the necessity - node, the special rules of coding, crossover and mutation are designed. Initial chromosomes are predisposed to improve the efficiency of algorithm. The steps of simulation to gain time and loss are presented. Finally, some experiments are performed and analyses are presented to verify the effectiveness of the algorithm.