Weathering is a significant process that alters the properties of microplastics (MPs) and consequently affects their environmental behaviors. In this study, we introduced a novel approach based on polarized light scattering technique, which offers advantages in terms of rapid, high-throughput, and submicron-sized detection. This technique was successfully applied to characterize the weathered MPs after a 180-day laboratory simulation of coastal environments. By employing polarization measurements, we obtained a 46-dimensional matrix data set for the weathered MP fragments and subsequently processed them using a backpropagation neural network. The successful extraction of effective polarization pulses confirmed the presence of MP fragments within the size range of 0.2-60 μm, yielding total accuracies for size classification ranging from 78.9 to 86.9%. Furthermore, this technique achieved an overall accuracy of 93.8% in classifying MPs with different weathering degrees and polymer types, revealing polarization parameters associated with size and morphological changes play a dominant role in characterizing the weathering process of MPs. Compared with conventional approaches, the novel polarized light scattering approach holds great promise for rapid, high-throughput, and accurate characterization of MPs with small sizes. The findings of this study provided new insights into how MPs change after long-term weathering in aquatic environments.