Abstract: Nuclear energy, as a low-carbon and highly efficient clean energy source, relies for its sustainable development not only on advanced reactor technologies and fuel cycle efficiency, but more critically, on the ability to safely and permanently dispose of the high-level radioactive waste(HLW) generated during its operation. The safe disposal of HLW constitutes the closing loop of the full life-cycle management of nuclear energy technology, serving as the ultimate safeguard for ensuring its environmentally friendly attributes. Whether this critical step is handled appropriately directly determines the social acceptance of nuclear energy as a “clean energy” label and fundamentally impacts its potential for long-term development. Therefore, the safe disposal of high-level waste is not merely a technical challenge; it is a core issue concerning energy strategy, environmental protection, and societal trust. The core challenges facing high-level waste disposal are multidimensional and extremely complex. Firstly, the complexity and uncertainty of radionuclide migration constitute the core scientific difficulty. The deep geological environment of a repository(e.g., groundwater chemistry, mineral composition, microbial activity, temperature field, stress field) and its evolution over ten-thousand-year timescales are fraught with variables, making the prediction of radionuclide(such as actinides and fission products) release, dissolution, migration pathways, and rates exceedingly difficult. Secondly, the long-term stability of the multi-barrier system is paramount. An ideal repository relies on a “defense-in-depth” strategy, combining an engineered barrier system(EBS)(e.g., waste matrix, waste container, buffer/backfill materials) and a natural barrier system(NBS)(i.e., the host geological rock). However, the degradation mechanisms of materials under high temperature, high radiation, chemical corrosion, mechanical stress, and over vast geological time scales remain a major research topic. Thirdly, the limitations of safety assessment over ten-thousand-year timescales present a unique challenge. Human experience and experimental data cannot directly cover such vast time spans. Safety assessments heavily depend on mathematical models and scenario simulations, the reliability of which hinges critically on the depth of understanding of key processes and the accuracy of model parameters, inevitably leading to extrapolation uncertainties and knowledge gaps. Finally, the socio-technical coordination challenge cannot be ignored. Even if a technical solution is scientifically feasible, socio-political factors such as public participation, risk communication, decision-making transparency, ethical considerations, and long-term stewardship responsibilities during siting, construction, operation, and closure are often as important as, or even more critical than the technical challenges. Specifically, the key process preventing the spread of radionuclides to the biosphere—radionuclide migration within the multi-barrier system—involves a series of coupled physical, chemical, and biological mechanisms. These include: radioactive decay(spontaneously reducing total nuclide inventory and activity), adsorption/desorption(retention and release of nuclides at solid-liquid interfaces, controlled by surface complexation, ion exchange, co-precipitation, etc.), advection(movement with bulk groundwater flow), matrix diffusion(molecular diffusion of nuclides in pore water or the material matrix, the dominant migration mechanism in low-permeability media), and fracture dispersion(diffusion of nuclides in rock fracture networks due to flow heterogeneity and hydrodynamic dispersion effects). The natural barrier(NBS), such as low-permeability clay rocks(e.g., bentonite) or crystalline rocks(e.g., granite), primarily retards nuclide migration significantly through its extremely low permeability, limiting groundwater flow velocity and flux, and the strong adsorption capacity(high distribution coefficient K_d) of mineral surfaces. The engineered barrier(EBS) provides active protection; for instance, high-integrity waste containers offer physical containment, preventing waste-water contact, while buffer/backfill materials(e.g., compacted bentonite) combine low permeability with strong chemical adsorption through their abundant clay minerals(e.g., montmorillonite) surfaces, further enhancing system safety. Significant progress has been achieved in the aforementioned key areas. Regarding radionuclide migration mechanisms, understanding has deepened concerning the chemical speciation, solubility, adsorption/desorption behavior, and colloid/nanoparticle-facilitated transport of key nuclides(e.g., neptunium, plutonium, iodine, technetium, selenium) in complex environments. In barrier material development, research continues on novel waste matrices(e.g., ceramics, glass ceramics), high-performance container alloys(e.g., copper, corrosion-resistant steels), and optimized buffer/backfill formulations(e.g., high-bentonite content mixtures), with long-term performance assessed through accelerated aging experiments. Safety assessment methodologies are also advancing, including more sophisticated coupled geochemical-transport models, probabilistic safety assessment(PSA) techniques, and validation based on geological analogues(e.g., the Oklo natural nuclear reactor) and natural tracers(e.g., Paleogroundwaters). This paper aims to provide a focused review of key aspects in high-level waste safety disposal research. We will delve into the chemical and physical mechanisms of radionuclide migration, analyzing key processes controlling migration rates and their interactions; systematically review the experimental methods and techniques used to study these mechanisms, including laboratory batch/column experiments, in-situ tests, advanced characterization techniques(synchrotron radiation, neutron scattering, high-resolution microscopy), and accelerated aging experiments; outline the mathematical models and numerical simulation methods employed for long-term behavior prediction, spanning coupled models from microscopic reaction kinetics to repository scale; and critically analyze the difficulties and technical challenges in current research, such as uncertainties in ultra-long-term prediction, experimental validation of barrier material behavior under extreme conditions, simulation of complex coupled processes(thermo-hydro-mechanical-chemical-biological, THMCB), and assessment of low-probability, high-consequence events. Finally, based on current knowledge and gaps, we will offer perspectives on future research directions, such as developing more intelligent in-situ monitoring technologies, exploring novel high-durability barrier materials, creating advanced methods for uncertainty quantification and model verification & validation(V&V), strengthening research on the integrated behavior of the geological disposal system(EBS-NBS coupling), deepening socio-technical systems integration studies, and promoting international collaboration and knowledge-sharing platforms.