Up to now, computational methods in chemistry and physics have been rapidly developed and matured, which make them useful tools in various fields of materials science. In this presentation, I will discuss about how the multiscale modeling methods consisting of quantum mechanics, molecular dynamics and mesoscale simulation methods, can make contributions to understanding given nanoscale systems and designing new systems. In the first part, I will present our recent progress for energy technologies such as fuel cell, battery, and solar cell, in which the nanometer-scale structures are modeled to understand/predict the atomistic-molecular mechanisms for various properties of interest, namely, ionic transport and redox potential via multiscale modeling. In the second part, I will also present modeling studies for various molecular assemblies such as block copolymer micelle, collagen self-assembly and so on. Here we will focus on how the nanometer-scale structures affect the properties of materials, especially the transport of small molecular species through the systems in order to accomplish desirable processes. Once the computational prediction is validated via comparing with experiment, the structure-property relationship can serve in designing new materials for better properties. In the last part of my talk, I will discuss the potential of machine learning briefly for high-throughput screening. The bottom line of my talk is that multiscale modeling methods can help achieve fundamental understanding of the nanoscale systems and establish design guide for new material development based on the structure-property relationship.