Abstract: Nitrogen-doped Diamond-Like Carbon thin films （N-DLC） can improve the frictional properties of part surfaces, and therefore, the research on the tribological properties of N-DLC thin films has been in high demand in recent years. Due to the limited amount of computational resources and computer running time, it is difficult to obtain a large amount of experimental data for studying the microscopic simulation of the interfacial structure evolution law in the friction experiments of N-DLC thin films. In order to explore the feasibility of cross use of molecular dynamics and artificial neural networks, and to gain a more comprehensive understanding of the friction properties and laws of N-DLC, BP neural networks, and KELM neural networks were referenced to the study of N-DLC. N-DLC is modeled using LAMMPS software, and molecular dynamics simulation data is used as the data source for artificial neural networks to train the two types of neural networks. Use validation samples to validate the two trained models, compare the prediction results of the two neural networks, and select the network model with the best performance. The results show that neural networks can predict the trend of hybrid bonds within N-DLC, with higher efficiency and less computational resources required. To a certain extent, it can replace the molecular dynamics simulation results to provide further analytical judgment. The study provides a useful exploration to promote the joint development of the two methods of molecular dynamics and artificial neural networks.