Aiming at the problem that the genetic algorithm(GA) is easy to fall into local optimum and the simulated annealing algorithm(SA) convergence rate is slow in the optimization of traveling salesman problem(TSP), a TSP optimization algorithm based on the improved genetic simulated annealing algorithm(IGSAA) is proposed. Firstly, the mathematical model is established according to the optimization goal. Then, the fitness function, crossover and mutation operators are improved in the part of genetic algorithm in order to make the algorithm more efficient to avoid falling into local optimum. Finally,an improved adaptive Metropolis criterion is proposed according to the evolution degree of each corresponding individual between old and new population in order to make the jump changing more adaptive in the part of simulation annealing algorithm which is propitious to the optimization of the algorithm. The experimental results on different TSP instances show that, the IGSAA algorithm designed can obtain better travel path on optimizing different TSP instances compared with other path optimization algorithms.