In this research, the problem addressed involves sets of practical case studies relating to a third-party logistics firm that mainly provides services to a big industrial estate in Thailand. This study considers constraints consisting of time windows, multiple trips, multi-product deliveries, a limited number of drivers and mixed fleets with limited and unlimited numbers of available vehicles. An adapted genetic algorithm hybridizes three inter-route search operators, i.e. relocation, exchange and elimination are developed and used to determine the best solution to a heterogeneous fleet vehicle routing problem with various constraints as mentioned previously. The benchmark problem sets do not exactly fit the addressed unique problem; therefore, the performance of the proposed method is evaluated against a branch-and-bound algorithm as a built-in solver. Due to the limitations of the solver for solving large-scale problems, percentage deviations of the solutions with respect to the best results and computation times are measured to confirm the quality of the presented method.