With the increasing focus on environmental issues, manufacturers are facing pressure to adopt end-of-life item collection practices and maximize the reuse of components or sub-assemblies to their fullest potential. The decision-making processes for disassembly planning and components inventory management become critical as they directly impact a company's production costs and service level. In this particular context, Reverse Material Requirements Planning (RMRP) is a widely utilized method to determine the optimal lot sizes of endof-life (EoL) products for disassembly, ensuring sufficient component supply. Traditionally, these lot sizes are calculated based on deterministic and dynamic demand assumptions when disassembly lead times are either zero or deterministic. However, in this work, we investigate optimization approaches for RMRP systems that consider disassembly lead time uncertainties, in a disassembly multi-echelon lot-sizing problem. We proposed stochastic programming models based on a Monte Carlo simulation approach whose performance is validated that are validated Their performance is validated on several numerical tests.