This study presented the Ant Colony System algorithm together with Local Search in solution of the fleet size and mix vehicle routing problem with time windows (FSMVRPTW). The purposes of this study were 1) to minimize the number of tours (or vehicles), and 2) to minimize the total travel distance where the objective of minimization of the number of tours takes precedence over the minimization of the total travel time. The FSMVRPTW is the planning of using various types of vehicles for the lowest cost of transportation which each type of vehicles have to load the limited goods and distribute them to the customers under the limited time. The researcher integrated the 2-Opt and Move-Exchange methods in the procedures to improve the initial solutions received from the Ant colony algorithm. The researchers employed the developed method to build up the algorithm through the use of Microsoft visual studio C++ version 6.0 with the 5 times of calculation for each problem data set. The parameter was set as a = 1.60, b = 0.1, and r = 0.1 respectively. The 15 Solomon problems including C1, R1, and RC1 sets were employed. The vehicles in this study were classified into two types namely 1) QA which has the maximum load for 200 units, and type QB having the maximum load for 100 units. The results revealed that the Ant Colony System algorithm together with Local Search gave the better solution that using the LINGO software developed by the researchers with the high quality of solution in general at the level of 1.47 percent.