Research efforts in molecular dynamics are dedicated to accurately determining equilibrium states of molecular systems, which are vital for predicting material properties such as melting points and thermal conductivity. Studies have evaluated that incorporating the three-body Axilrod-Teller potential into a two-body model increases simulation accuracy. AutoPas is an open-source library that automatically selects the best-fitting strategies for any given particle simulation scenario, aiming to deliver optimal node-level performance. The linked cells algorithm, used to efficiently identify neighboring particles, is part of this selection pool. There are multiple cell traversal schemes in AutoPas for pairwise interactions. This thesis implemented three-body linked cell traversals with the C08 base step for three-body interactions at the core, and we extended the implementational idea of the C08 base step to N-body interactions. A performance analysis of three-body traversals across simulations of different density and cell size configurations showed the dominance of C08 among all other traversals, including C08-based traversals and the C01 traversal. Cell Size Factor of 0.5 achieved the best performance in most simulations. Besides, the study highlights linked cell limitations due to the low hit rate for three-body interaction. The evaluation examines the balance point of performance gains by hit rate increase and the price of administrative overhead of cells for trying to achieve higher hit rates. This overhead constraint limits the use case of linked cells for three-body interactions to hybrid algorithms that use spatial information, e.g., Verlet lists, and to highly dense systems, where the proportion of force computation dilutes the cell overhead.