Disassembly scheduling is one of the important problems in reverse logistic decisions. This paper focuses on this problem with capacity restrictions on disassembly resources, lost sales, multiple products and without part commonality. A model with two objectives is developed and optimized by a multi-objective approach. The first objective is a sum of several costs to minimize : setup cost, inventory cost, and over capacity penalty cost. The second objective is a measure of the service level. Considering the complexity of this model, a genetic algorithm is developed (NSGA-II) to obtain a set of Pareto-optimal solutions, the results are compared with those calculated by a mixed integer programming model. Results of computational experiments on randomly generated test instances indicates that the genetic algorithm gives good quality solutions up to all problem sizes in a reasonable amount of computation time whereas linear programming solvers do not give solution in reasonable time.