Lost　circulation,　as　a　widespread　and　costly　drilling　problem,　not　only　wastes　a　lot　of　drilling　fluid,　but　also　produces　a　large　amount　of　nonproductive　time　(NPT).　Therefore,　detecting　the　symptoms　of　lost　circulation　opportunely,　pinpointing　the　loss　depth　and　evaluating　the　loss　rate　precisely　are　extremely　crucial.　A　diagnosis　method　for　detecting　the　occurrence　of　lost　circulation　and　identifying　the　loss　depth　as　well　as　loss　rate　at　an　early　stage　is　proposed,　which　combines　a　transient　pressure　and　temperature　coupling　model　with　unscented　Kalman　filter　(UKF)　estimation.　We　first　establish　the　transient　pressure　and　temperature　coupling　model　under　the　conditions　of　normal　drilling　operation　and　lost　circulation.　Then　the　detection　and　identification　estimators　for　estimating　pressure-loss　factors,　flow　rate　factor,　loss　depth　and　loss　rate　are　constituted　by　embedding　the　detection　and　identification　models,　which　are　derived　from　the　transient　pressure　and　temperature　coupling　model,　into　UKF　estimation.　In　the　end,　a　simulated　case　study　is　presented　to　illustrate　the　performance　of　the　diagnosis　method.　In　the　simulation,　the　estimated　pressure　and　outlet　flow　rate　trace　the　measurements　well　with　the　evolution　of　pressure-loss　factors　and　flow　rate　factor.　Moreover,　the　pressure-loss　factor　in　annulus　and　flow　rate　factor,　which　fluctuate　to　a　little　extent　in　normal　drilling　condition,　deviate　from　the　base　value　once　lost　circulation　occurs.　The　simulation　indicates　the　superior　performance　of　the　diagnosis　method　with　the　introduction　of　pressure-loss　factors　and　flow　rate　factor,　and　the　average　error　of　estimated　loss　depth　is　less　than　5%.　In　addition,　the　parametric　sensitivity　analysis　illustrates　the　universality　of　the　diagnosis　method,　which　has　certain　ability　to　improve　drilling　safety　and　efficiency　with　lower　NPT.