In most countries, the preferred long-term storage option for high level waste (HLW), is encapsulation followed by disposal in geological repositories at depths of hundreds of metres. One of the safeguards measures in the construction and operational phase is to verify that the as-built facility, including the above and below ground facilities, is in agreement with declared repository design information. Thus, it is vital to further enhance methods for design information verification (DIV). In particular, the presence of unknown voids or tunnels in the surrounding geology could threaten a repository’s integrity, in terms of both environmental safety and nuclear safeguards. Ensuring continuity of knowledge (CoK) for all nuclear material in the process of transport, encapsulation, and disposal in a geological repository is a key element of the safeguards approach. Different technologies have been discussed for DIV of an encapsulation plant or geological repository, such as 3D laser scanning, simultaneous location and mapping (SLAM), geophysical monitoring, surveillance cameras, and satellite imagery. Here, we introduce muon radiography as another, probably complementary potential technology. Muon radiography is an imaging technique that exploits the attenuation interactions of cosmic ray muons in matter. Muons are attenuated to a greater degree in denser material, and so the presence of voids or other geological features in a large dense structure can be deduced from excesses or deficits of muons in the corresponding directions. Muon radiography therefore represents a possible method for identifying voids, tunnels or other unknown features in and around a geological repository. We describe a Monte Carlo simulation study to determine the potential of muon radiography for design information verification of a geological repository for HLW. A detailed digital model of a geological repository has been designed, with particle detectors inserted and new features such as voids added. The limits of the method in terms of the minimum identifiable feature size within practical exposure times are established, and the effects of detector placement and design are interrogated.