In　this　paper,　the　ultrasonic　frequency　shift　of　in-vitro　porcine　liver　with　temperature　change　during　the　water　bath　heating　process　was　analyzed　and　studied　using　the　Autoregressive　(AR)　model,　based　on　the　discrete　random　medium　model　of　biological　tissues.　The　power　spectrum　density　of　the　ultrasonic　backscattered　signal　was　computed　by　establishing　the　AR　model.　Statistical　analysis　was　performed　on　the　experimental　results.　The　experimental　results　show　that　the　mean　frequencies　of　ultrasonic　backscattered　signals　of　the　porcine　liver　tissues　change　with　the　temperature　variation　from　20°C　to　60°C.　The　mean　frequency　is　(3.55　±　0.09)　MHz　at　20°C,　(3.14　±　0.12)　MHz　at　30°C,　(2.87　±　0.11)　MHz　at　40°C,　(2.79　±　0.10)　MHz　at　50°C,　and　(2.72　±　0.08)　MHz　at　60°C.　These　data　suggest　that　the　mean　frequency　decreases　as　the　temperature　increases.　The　mean　frequency　obtained　using　the　AR　model　method　may　be　used　as　a　characteristic　parameter　for　tissue　characterization　in　the　recognition　of　coagulation　alterations　during　microwave　hyperthermia,　which　may　offer　a　potential　pragmatic　prospect　for　the　practice　of　noninvasive　monitoring　of　microwave　hyperthermia.