The work aims to study the inventory routing optimization of a two-level supply chain distribution system consisting of a supplier and multiple retailers based on the vendor-managed inventory model, determine the length and quantity of delivery cycle in the planning period, coordinate and optimize daily sales volume to inventory strategies and distribution plans in the supply chain, so as to minimize the total cost of the supply chain system. The costs considered included inventory holding cost, out of stock cost and distribution cost. In order to minimize the total cost of the supply chain system, a new strategy with variable cycle was proposed. The genetic algorithm was used to solve the optimal scheme. The results of different strategies were compared and analyzed. The results showed that, compared with the fixed cycle strategy, in the variable cycle strategy, the lowest saving ratio of the total system cost was 1.7% and the highest saving ratio was 42.3%. The total cost of the system can be effectively saved by dividing and adjusting the length and quantity of each distribution cycle in the planning period. Meanwhile, the distribution scheme with multiple vehicle types is obviously better than that with the same vehicle types.