Abstract: In practice, it is common for retailers to bundle the products from multiple manufacturers. For example, Gome launches "complete sets of home appliance", which bundles several well-known brands of TVs and refrigerators. These products are provided by multiple upstream manufacturers. Ctrip.com bundles air transportation, hotel accommodation, etc., are also from different suppliers. Based on the above business phenomenon. A three-stage game model for members in a supply chain system consisting of two manufacturers and one retailer is constructed in this paper. The manufacturers invest in R&D to reduce the cost of their products and decide the wholesale price of products. The retailer then orders the product from the manufacturers and selects the appropriate sales method, i.e., separate sales or bundled sales. By comparing the R&D level of the manufacturers and the profit of the supply chain members between separate sales and bundled sales under different product relevance, the supply chain member's preference for different sales methods is explored. It is found that, for independent products, separate sales is always the dominant strategy, but there is no common preference among system members; for complementary products, the role of non-coordination at the supply side of the product and the differences in perceived value of consumers will exceed the complementary linkage advantage of the product itself, resulting in the failure of the binding strategy and making the separate sale as a common preference for system members; for substitute products, the low cost products with the moderate consumer perception differences and the high cost products with less consumer perception differences are more profitable under bundling. The research results of this paper provide good theoretical guidance for the practical application of the bundled sales strategy.

Key words: supply chain, product relevance, cost R&D, bundled sales, separate sales