Particle shape analysis is essential in sedimentological research, as it offers vital insights into the sedimentary environment and transport history. However, little is known about the particle shape variation across different sand fractions, as well as the differences between particle shape data based on volume and number weighting. In this study, we investigate the grain shape variation of different sand-sized particles (fine, medium, and coarse sand fractions) in aeolian dune (11 samples) and lake beach (12 samples) environments around Poyang Lake, China, using dynamic image analysis (DIA). The shape data results based on both volume-weighted and number-weighted methods reveal significant differences in shape parameters (circularity, symmetry, aspect ratio, and convexity) among different sand fractions, especially between coarse and fine sand. This highlights the critical need for size-fractionated analysis when employing particle shape as an environmental discriminant. By integrating 86 sets of published particle shape data from different depositional environments, we found that volume-weighted shape data has limited ability to differentiate beach and dune sands, although it distinguished the fluvial, desert dune, and coastal beach sand well. In contrast, number-weighted shape data effectively distinguished the beach and dune sands, as fine sand particles are typically transported in suspension during fluvial processes and in saltation during aeolian processes. This demonstrates the role of integrating both volume-weighted and number-weighted shape data in future studies to accurately distinguish sedimentary environments.