Traveling Salesman Problem (TSP) is an NP-hard optimization problem that can be solved by a heuristic composite algorithm. A composite algorithm is a heuristic optimization model that combine tour construction algorithm and tour improvement algorithm. Clarke Wright Savings heuristic is one of the best methods that produce a good initial solution, and local search is known to be a successful operator to make an improvement solution. This paper will present a composite algorithm as a preliminary model based on Clarke wright savings and local search K-opt to solve TSP. The experimental result shows that the proposed algorithm can solve a large problem instance of Traveling Salesman Problem up to 85.900 points, with competitive results, small variations of computing time for 30 problem instances, and relatively short computing time.