Abstract: An autonomous relative navigation method based on sequential image measurements of unknown landmarks for planetary landing in unknown environment is presented. In the proposed method, the line-of-sight vectors of landmarks with unknown 3D positions tracked in three sequence images and position information derived from inertial measurement unit at the time instances of the three images are used to develop implicit measurement constraints. An implicit extended Kalman filter tightly integrates the constraints with the measurements from IMU to estimate the lander's position and velocity relative to the landing site. Numerical simulations of Mars landing indicate the proposed navigation method is efficient and meets the relative navigation requirement of safe planetary landing.