Abstract:The end-of-life (EOL) photovoltaic (PV) modules are becoming an important secondary resource stream, and their safe, efficient recycling strongly depends on the performance of sorting technologies during the pretreatment stage. This review examines sorting technologies for EOL PV modules with the aim of clarifying the roles, application ranges, and limitations of traditional physical sorting and intelligent sorting, and identifying technical directions for improving recycling efficiency and reducing secondary pollution. Based on recent research and engineering practice reports, the paper classifies current sorting routes into traditional physical processes—such as high-pressure electrostatic separation, eddy current separation, magnetic separation, and gravity separation—and intelligent sorting systems driven by machine vision, deep learning algorithms, and precision positioning equipment. The literature is synthesized to compare these routes in terms of separable material types, particle-size ranges, dependence on manual parameter adjustment, and the distinguishability of materials with subtle differences in properties such as conductivity, density, or surface characteristics. Reported data on recovery and purity of product streams, operating stability, and control complexity are used to summarize the typical roles of different sorting technologies within complete EOL PV recycling flowsheets. The results of the review indicate that traditional physical sorting is suitable for the basic separation of glass and metallic fractions and has advantages in process simplicity and robustness, but it is generally restricted to fragments in the range of 2–20 mm and to systems in which materials exhibit pronounced differences in physical properties. These routes have limited capacity to deal with laminated structures and components with similar compositions, and they usually require frequent manual tuning to maintain stable recovery and purity. Intelligent sorting technologies can identify wafers, glass, ribbons, and back sheets at the single-particle level by analyzing intrinsic optical and morphological features, expanding the applicable size range and reducing reliance on manual operation. Studies further suggest that coupling intelligent recognition and actuation modules with electrostatic, magnetic, or gravity separation units improves overall separation precision and decreases the risk that hazardous or high-value components may enter inappropriate product streams. From the comparative analysis, the main technical bottlenecks are identified as the narrow applicability and low adaptability of traditional physical processes, along with the high equipment cost, the requirement for large, high-quality datasets, and system integration challenges associated with intelligent sorting. The review concludes that future development should focus on upgrading conventional lines through the integration of intelligent perception and control, designing modular intelligent sorting units that can be flexibly combined with different pretreatment and separation processes, and coordinating technological innovation with policy measures and standardization. These directions are expected to support higher-efficiency, lower-pollution sorting systems and promote the green and high-value utilization of EOL PV modules. Close-