Establishing a robust facility location and assignment plan to improve the efficiency of the decontamination process is critical to alleviating the physical impact of the radiation leakage that occurs in a nuclear accident. This study develops an approach for optimizing the locations of decontamination facilities and assignments of affected villages. The approach is a robust optimization model that optimizes the worst-case performance. The system dynamic model is integrated into the robust optimization model to simulate the decontamination process and compute the decontamination time. A case study is conducted of the Plume Emergency Planning Zone in China. The results indicate that (1) a decontamination site location plan can be obtained in which each site is located in a different direction, (2) no evacuee will be allowed to travel across the downwind area in an assignment plan, and (3) a larger financial investment does not imply an increased decontamination efficiency. An appropriate budget exists that can balance the decontamination time and cost. The proposed model can assist decision makers in (i) better understanding the effects of decontamination site location and village assignment and (ii) deciding which location and assignment plans should be applied to cope with disruptive nuclear accidents.