For the shortcomings of greed evolutionary algorithm (GEA) in solving the traveling salesman problem, such as smallscale, low success rate, this paper induces Inverover down different operators, and greedy operators, improves close neighbor optimization, mapping operator, mutation operator strategy, proposes a new and improved evolutionary algorithm to solve the mediumscale traveling salesman problem (TSP). Simulation results verify that the evolutionary algorithm converges fast, and is of the advantages of a high success rate, better stability.