End-of-life (EOL) product recycling has received increasing attention because of potential environmental, social and economic benefits. A well-designed reverse supply chain (RSC) can efficiently handle EOL products. As the critical activity in the RSC, the disassembly process decomposes collected EOL products into components to fulfill the demands of remanufacturing plants. Efficiently coordinating RSC design and disassembly line balancing decisions may improve the whole system performance significantly, especially when facing multiple EOL products and uncertainty. This paper investigates a novel integrated RSC design and disassembly line balancing problem to handle multiple EOL products where the supply of EOL products, the demand for components, and task times in disassembly are stochastic. This complex problem needs to jointly determine the number and locations of disassembly plants, disassembly equipment procurement, disassembly line balancing, and inventory levels of both EOL products collected and components dismantled. The objectives are to maximize the expected profit and minimize the carbon emissions simultaneously. For the problem, a bi-objective two-stage stochastic programming model is formulated and an exact ɛ-constrained method is proposed, transforming the bi-objective problem into a series of single-objective problems. Especially, an improved Benders decomposition approach is developed to solve each single-objective problem efficiently. Numerical experiments comprising an illustrative case and 200 random instances are conducted to evaluate the performance of proposed methods. Moreover, some managerial insights are drawn.