Abstract Accurate identification of cloud thermodynamic phases ‐warm water, supercooled water and ice—is essential for the global energy budget and hydrological cycle, and serves as a crucial prerequisite for retrieving cloud optical and microphysical parameters. Existing methods each have inherent limitations, and relying on a single approach often leads to ambiguous classifications. In this study, we develop a collaborative algorithm that integrates shortwave infrared, water vapor absorption and polarization measurements (hereafter referred to as SWaP) to retrieve cloud thermodynamic phases, relying solely on data from the Shortwave Infrared Polarization Multi‐Angle Imager onboard the Fengyun‐3G satellite. The SWaP algorithm is applied to a typhoon case, and the resulting spatial distribution of warm water, supercooled water, and ice clouds shows reasonable structure. The polarization signatures and brightness temperature characteristics of the retrieved cloud phases are highly consistent with theoretical expectations, demonstrating the reliability of the proposed algorithm.