Location-routing problem (LRP) is a combinational optimization problem in a logistics system. Most heuristic methods employed for LRP is dividing the problem into location assignment and vehicle routing with a two-phase method, but this method often does not lead to a satisfactory result for the information can not be compressed from one phase to the other efficiently. In this paper we are concerned with a particular type of facility location problem in which there exists two echelons of facilities, and use a hybrid genetic algorithm to solve the problem regarding its solution as a whole, and thus easily obtain the optimal solution. Genetic algorithm uses a three-level chromosome coding method composed of binary code and integer code. Improving crossover and mutation with tabu search algorithm to increase search efficiency. The numerical results indicate that the method is feasible for practical applications.