In the Traveling Salesman Problem, the goal is to find the shortest route for a salesman who wants to pass through N cities in such a way that she/he passes through all the cities exactly once and returns to the starting city. In this project, genetic algorithm is used to solve TSP problem. To evaluate the model, the data of three sets 535ali, 2103d and 29baysg from TSPLIB have been used. This data including necessary explanations are available via this link: https://www.iwr.uni-heidelberg.de/groups/comopt/software/TSPLIB95/tsp For training part, competitive selection method, rotary recombination method and insertion mutation method are used and the initial population is randomly selected, also the value of mutation rate is 2% and recombination rate is 90%. To solve the problem in asymmetric mode (TSP Asymmetric), ftv33 data with 33 cities have been considered. In this case, the distance (cost) of going from city i to city j is different from the opposite distance (cost) of going from city j to city i.