The Fukushima accident severely damaged fuels in three nuclear reactors. Even though part of information on their condition has been identified by investigations with advanced methods, the condition of the fuel is still uncertain. We developed a model to simulate the uncertain condition of damaged fuel debris and evaluated its nuclear radiation. Our research is focused on the uncertain geometry of fuel debris in a canister, and the uncertain condition is modeled by random sampling on size and distribution of the fuel debris in a canister. A Monte Carlo based module of SCALE code is used for the calculation of flux and dose rate, and a Python code is made for the random sampling and the evaluation of uncertainty. The uncertainty of dose rate by the uncertain geometry of fuel debris in a canister is evaluated as the coefficient of variation and the percent relative range of the dose rate.