With the development of the world economy, more and more pavements are constructed to guarantee the functions of logistics and tourism etc. among different regions. However, annual costs to keep the pavement in a good state are huge every year, including maintenance, repair and reconstruction of pavement. So it is necessary to improve the properties and extend the service life of pavement, in order to save huge financial expenses. The properties and service life of asphalt pavement is caused by many factors, of which the properties of asphalt binder are very important, so some scholars have researched the properties of asphalt binder. However, the normal research approaches are some macro experiments, it is not clear that the mechanism of the properties of asphalt binder. So a Molecular Dynamics (MD) simulation is employed in this thesis to study the asphalt binder, because it is a sufficient approach to understand asphalt binder from a micro perspective. Firstly, four kinds of asphalt binders with different stiffness are treated by Rolling Thin Film Oven Test (RTFOT) and the Pressure Aging Vessel (PAV) test, obtaining twelve asphalt binders at three different aging states. Then some mechanical properties tests including the penetration, ductility, softening point, Dynamic Shear Rheometer (DSR), Bending Beam Rheometer (BBR) are employed to characterize their mechanical properties at different temperature, meanwhile, a novel six-fraction (saturate, monoaromatics, diaromatics, polyaromatics, resin and asphaltene) separation is also used to clarify the fractional composition of these twelve asphalt binders. Afterwards the relation between fraction, mechanical properties and aging states are analyzed.After understanding the relation between the fractions and properties of asphalt binder, the mechanism of the influence of polyaromatics on the properties are explored using MD simulation at a micro scale. Results show that the polyaromatics cause the cohesive energy density to decrease and the ratio of free volume in asphalt binder to increase, which is the micro mechanism of its softening effect. Then, two molecular models of asphalt binder with different paraffin contents are created to study the effect of paraffin on the mechanical properties of asphalt binder. Results show that the Young's modulus decreases while the Poisson's ratio increases with the paraffin content. The bulk modulus and shear modulus both decrease with increasing paraffin content. In addition, the diffusion coefficient of the neat asphalt binder model has a very high dependency on the temperature. The self-healing behavior of asphalt binders with different paraffin contents is slower than the asphalt binder without paraffin, i.e. the paraffin prevents the self-healing behavior of asphalt binder in a certain degree. The use of a co-production of renewable bio-oil as a modifier for petroleum asphalt is recently getting more attention in the pavement field due to its renewability and the optimization for the conventional petroleum-based asphalt binder. Significant research efforts have been done which mainly focused on the mechanical properties of bio-asphalt binder. However, there is still a lack of studies describing the effects of the co-production on performance of asphalt binders from a micro scale perspective to better understand the fundamental modification mechanism. Therefore, a reasonable molecular structure for the co-production of renewable bio-oils is created based on previous research findings and the observed functional groups from the Fourier-transform infrared (FTIR) spectroscopy tests, which are fundamental and critical for establishing the molecular model of the bio-asphalt binder with various biomaterials contents. Molecular simulation shows that the increase of biomaterials content causes the decrease of cohesion energy density, which can be related to the observed decrease of dynamic modulus. Additionally, a parameter of Flexibility Index is employed to characterize the ability of asphalt binder to resist deformation under oscillatory loading accurately. The aging of asphalt binder is a vital part of the pavement’s service life and has therefore been subjected to a lot of research regarding the chemical composition and rheological properties of asphalt binder. However, the microscopic reasons for these changes are not yet clear. A molecular dynamics simulation is employed to investigate the aging process from a microscale perspective. Six molecular models for six kinds of asphalt binders at different aging states are created according to the ratio of saturates, aromatics, resin and asphaltene (SARA) obtained through fractional separation process as well as carbonyl and sulfoxide indexes obtained from FTIR spectroscopy. Some thermodynamic properties including density, cohesive energy density, surface free energy and adhesion behavior of the asphalt binders are simulated, after which some correlations among them are explored in more detail. The results show that, aging causes the ratio of free volume in the model to decrease. Furthermore, the increase of oxygen causes an increase of molecular polarity, resulting in stronger intermolecular attraction forces indicating an increase of cohesion, which raises the complex modulus. Lastly, aging causes a decrease of surface free energy for asphalt binder, which indicates that asphalt binder has a downward tendency of adhesion. In other words, the reason of the deterioration of adhesion property is the decrease of surface free energy. Finally, two asphalt binder models are created based on the conventional SARA method and the novel proposed six-fraction method to propose a more accurate model with the asphalt binder used in this thesis. In order to verify the accuracy of the two molecular models, the Atom Force Microscopy (AFM) experiment is conducted. A comparison of the results shows that, comparing to the SARA model, the novel six-fraction model can better characterize the density and elastic modulus of the asphalt binder, indicating the six-fraction asphalt binder model is more accurate and can help us to better understand the micro-properties. Based on this novel six-fraction model, some research in depth will be further conducted in the near future.