Logistics distribution path optimization as an NP-hard problem is one of the important problems in logistics. Many intelligent algorithms are considered to be used to solve such problems. In this paper, an improved genetic algorithm is proposed for solving Travelling Salesman Problem (TSP), a kind of classical logistics distribution path optimization problem. The method uses island model genetic algorithm to formulate rules that are more suitable for TSP, through applying greedy algorithm in the generation of initial population, modifying the selection method and discussing different migration strategies, the computation time of solving the problem is shorten, the calculation efficiency is improved, and the probability of falling into the local optimal solution is reduced. Finally, experiments are conducted to discuss the effectiveness of our method.