<p>Shape-based separation of biological particles has elucidated cellular morphology and function. However, length-based sorting of helical-shaped microorganisms, such as <i>Arthrospira</i> (<i>Spirulina</i>) <i>platensis</i>, in microfluidic platforms has not yet been achieved. Here, we report a passive, length-based sorting method for <i>A. platensis</i> trichomes using a straight microchannel that exploits a self-alignment effect of helical filaments. In our design, trichomes flowed through a straight microchannel narrower than their average length and entered a downstream expanding channel with five outlets. High-speed camera-based observation revealed that shorter trichomes predominantly aligned near the channel walls, whereas longer trichomes spanned the channel width. These alignment tendencies, termed the self-alignment effect, were flow-rate dependent and most prominent at a Reynolds number of 40. Thus, short and long trichomes were guided to different outlet positions. Quantitative analysis showed that at a threshold length of 300 µm, sorting purities of 77% and 84% were achieved for short and long trichomes, respectively. This demonstrates effective, label-free separation of multicellular cyanobacterial filaments based on length in a simple microfluidic geometry. As trichome length reflects the growth stage and environmental response, this method offers a promising platform for studying physiological heterogeneity in <i>A. platensis</i> populations. Furthermore, it can facilitate applications in biotechnology and materials science, where precise selection of filamentous morphologies is desired. This principle can also be applied to other microorganisms or flexible helical structures. The observed self-alignment effect may become a new paradigm in passive microscale sorting.</p><p></p>