In order to solve the capacitated vehicle routing problem(CVRP), a hybrid variable neighborhood symbiotic organisms search(HVNSOS) algorithm is proposed. A new encoding scheme called ordered encoding is designed, which connects customers' priority sequence with vehicle reference points. Based on the ordered encoding, the SOS algorithm can be applied to discrete optimization problems(e.g. CVRP). For the sake of enhancing the global searching ability of the algorithm, three symbiotic operators are constructed according to the features of the encoding scheme. The symbiotic operators can also expand the search space of the CVRP. To improve the quality of solutions, the algorithm combines the variable neighborhood search algorithm to design three new local search operators of relocate, exchange and 2-OPT. The comparison and analysis of the results show that the proposed algorithm can solve the CVRP effectively, and it is superior in the performance of computing efficiency and global stability to the algorithms compared.