* S Supporting Information CONSPECTUS: Self-assembling structures and their dynamical processes in polymeric systems have been investigated using three-dimensional transmission electron microscopy (3D-TEM). Block copolymers (BCPs) self-assemble into nanoscale periodic structures called microphase-separated structures, a deep understanding of which is important for creating nanomaterials with superior physical properties, such as high-performance membranes with well-defined pore size and high-density data storage media. Because microphase-separated structures have become increasingly complicated with advances in precision polymerization, characterizing these complex morphologies is becoming increasingly difficult. Thus, microscopes capable of obtaining 3D images are required. In this article, we demonstrate that 3D-TEM is an essential tool for studying BCP nanostructures, especially those self-assembled during dynamical processes and under confined conditions. The first example is a dynamical process called order−order transitions (OOTs). Upon changing temperature or pressure or applying an external field, such as a shear flow or electric field, BCP nanostructures transform from one type of structure to another. The OOTs are examined by freezing the specimens in the middle of the OOT and then observing the boundary structures between the preexisting and newly formed nanostructures in three-dimensions. In an OOT between the bicontinuous double gyroid and hexagonally packed cylindrical structures, two different types of epitaxial phase transition paths are found. Interestingly, the paths depend on the direction of the OOT. The second example is BCP self-assemblies under confinement that have been examined by 3D-TEM. A variety of intriguing and very complicated 3D morphologies can be formed even from the BCPs that self-assemble into simple nanostructures, such as lamellar and cylindrical structures in the bulk (in free space). Although 3D-TEM is becoming more frequently used for detailed morphological investigations, it is generally used to study static nanostructures. Although OOTs are dynamical processes, the actual experiment is done in the static state, through a detailed morphological study of a snapshot taken during the OOT. Developing time-dependent nanoscale 3D imaging has become a hot topic. Here, the two main problems preventing the development of in situ electron tomography for polymer materials are addressed. First, the staining protocol often used to enhance contrast for electrons is replaced by a new contrast enhancement based on chemical differences between polymers. In this case, no staining is necessary. Second, a new 3D reconstruction algorithm allows us to obtain a high-contrast, quantitative 3D image from fewer projections than is required for the conventional algorithm to achieve similar contrast, reducing the number of projections and thus the electron beam dose. Combining these two new developments is expected to open new doors to 3D in situ real-time structural observation of polymer materials.