Abstract: The principle of information entropy is applied to seafloor terrain matching navigation, and the terrain variance entropy is established. The matching algorithm including the search phase and the positioning phase is designed. The position error accumulated over time by inertial navigation system (INS) is corrected through coarse matching and fine matching. Further, the actual seafloor data is used for simulation. Results show that the position matching of longitude and latitude has a fast convergence speed, it can effectively correct the position error of inertial navigation system. This method has the advantages of high accuracy and good stability, and has certain practical value.