In a reverse logistics context, this study addresses a multi-period disassembly lot-sizing problem under uncertainty disassembly lead times. The disassembly lead time comprises the time elapsed between releasing the disassembly order and receiving the disassembled item. To the best of our knowledge, this uncertainty is studied for the first time in the disassembly lot-sizing problem. A disassembly system with a two-level BOM and a single end-of-life product type is considered. Disassembly lead times are discrete random variables with a finite number of possible values. The problem is formulated as a two-stage stochastic Integer Linear Programming (ILP) model through all possible scenarios in order to minimize the expected total cost. Because of the large number of scenarios, ILP is intractable. To make it treatable, we propose an optimization approach that combines Monte Carlo (MC) simulation and ILP. Besides, to solve large scale problems, we propose a basic genetic algorithm. Experimental results based on randomly generated data show the effectiveness of the proposed.