Nowadays, due to irregular human intervention and exploitation of available resources, the environmental crisis has been one of the biggest threats in the world. These risks can sometimes come from destroying habitats and arable land. According to the importance of this issue, this study tries to design a construction and demolition wastes (C&DW) reverse logistics network. For this purpose, a bi-level mixed-integer linear programing (MILP) model is designed. The lower level determines location of C&DW recycling and disposal centers and the upper-level involves allocation of waste to the recycling and disposal centers. A scenario-based robust optimization approach is applied to cope with the uncertainty in the percentage of C&DW transported from the generation points. Eventually, the final model is solved in GAMS optimization software and its performance and efficiency are analyzed. Mazandaran province, Iran, is selected as case study because of its geographical conditions. The results show that allocation and establishment decisions are made in a way that impose the least cost to the network in different scenarios. In addition, the scenario-based robust optimization approach represents less sensitivity to solutions for considered scenarios. Also, considering the hierarchy model leads to appropriate solutions with less costs compared to the separate leader and follower models.