At　present,　among　the　known　methods　of　extracting　interferometric　phase,　some　methods　developed　from　spatial　spectrum　estimation　and　extracting　interferometric　phase　by　MUSIC　or　TLS-ESPRIT　algorithm　are　being　perfected.　However　these　methods　can　bring　on　confusion　of　different　scatter　centers　among　Interferometric　phase　due　to　the　lack　of　measures　of　mechanism＇s　separation.　By　calculating　the　power　of　backscatter　signals,　the　TEL-ESPRIT　algorithm　for　joint　estimation　of　interferometric　phase　and　power　of　backscatter　signals　is　proposed,　TLS-ESPRIT　algorithm　can　classify　phase　by　power　and　achieve　separation　of　scatter　centers.　Therefore　it　can　improve　estimate　precision　of　vegetation　height.　In　this　paper,　the　new　algorithm　is　described　in　detail.　Using　SIR-C/X-SAR　L　band　data,　its　effectiveness　is　tested.