Traveling salesman problem (TSP) is of important theoretical and practical significance and applied widely in engineering practice. The genetic algorithm, ant colony algorithm and simulated annealing were adopted to solve the traveling salesman problem, and the Chinese traveling salesman problem was chosen to simulate. Through the comparison of these three algorithms' advantages and disadvantages, their different applications were gained: the ant colony algorithm is suitable for slow and accurate solving, the simulated annealing applies to quick and accurate solving, and the genetic algorithm is for quick but low accurate solving.