In warehouse management order picking is one of the key operations that incur high costs as well as one of the most complex tasks. It comprises the construction of order batches, batch assignment, picker routes, and scheduling of pickers. Therefore, the development of an efficient order picking system and the optimization of these operations have significant effects on the overall efficiency of the warehouse. This paper focuses on studying and modeling the order batching, batch assignment, and picker routing problems in a multi-warehouse, multi-period, multi-picker order picking system. We propose a multi-objective mathematical model for minimizing the delivery times of batches and the total cost of order picking operations. Also, for the first time, a possibilistic approach is applied to overcome uncertain conditions in the order picking problem. Given the complexity of the problem, Benders' decomposition is implemented to solve the proposed model. The applicability of the proposed method is evaluated through a range of small to large test problems and an actual case study. The results indicate that the proposed exact method is capable of finding high-quality solutions within a reasonable computational time and number of iterations, which serves as evidence of its suitability for large-scale, complex real-world industrial contexts.