Traditional　polarimetric　current-sensing　methods　contain　many　optical　devices,　leading　to　poor　reliability.　Also,　their　sensing　output　is　distorted　by　massive　system　noise.　To　address　these　issues,　this　paper　presents　a　polarimetric　current-sensing　method　using　single-axis　polarization　encoding　(SPE)　and　a　dynamically　decoding　algorithm.　A　single-axis　circulator　and　a　Faraday　rotator　mirror　are　used　to　encode　the　measured　current　by　light.　The　SPE　only　requires　these　two　optical　devices　in　the　light-propagation　path,　enhancing　the　measurement　reliability.　The　encoding　process　can　significantly　suppress　the　noise　by　eliminating　high-order　terms.　To　correctly　restore　the　current　signals,　the　light　intensity　is　decoded　by　Jones-matrix　analysis　and　dynamic　parameters.　Finally,　normal　currents,　fault　currents　and　harmonic　currents　in　grids　are　used　to　test　the　performance　of　the　presented　method.　The　results　show　that　the　maximum　amplitude　error　is　smaller　than　1.3%　in　a　wide　range　of　amplitudes,　frequencies,　temperatures　and　vibrations,　and　the　signal-to-noise　ratio　and　number　of　devices　required　in　the　path　are　51.43　dB　and　2,　respectively　(those　of　the　recently　published　polarimetric　method　are　1.7%,　-4.24　dB　and　5,　respectively).　IEEE