The main goal of designing magnetic fusion device structures is to imitate the interactions that occur between plasma and the surfaces of structural materials. The sputtering effects of plasma ions on the surface structures of materials are not the same for all materials. We can gain a lot of information about the inter-atomic potential energy of these structures by using molecular dynamics and atomic simulation configurations. It is important to evaluate a range of inter-atomic potential energy algorithms while analyzing the crystal structure of material surfaces. This work used different kinds of machine learning approaches to estimate inter-atomic potential in order to increase the precision of the surfaces of bulk crystal materials that interact with plasma ions. This will enhance the structural design of magnetic nuclear fusion reactors by improving the precision of molecular dynamic simulations.