In order to solve the Distance-constrained Capacitated Vehicle Routing Problem ( Distance-constrained CVRP ), the cellular genetic algorithm(CGA) is used in this paper. A new crossover operator--SAX which can better reflect the self-adaptability of CGA is proposed, and three types of neighborhood are introduced to analyse the search performance of CGA in Distance- constrained CVRP. Two instances are introduced to show the feasibility of CGA in solving Distance-constrained CVRP. The experimental results show that the performance of CGA is obviously better than that of the traditional genetic algorithm, and the optimization results of the vehicle routings are improved. Especially, the Moore neighbor structure indicates better search efficiency for SAX crossover operator. Because of the strong searching ability, CGA can effectively solve the optimization of vehicle routing problems.