The trajectory guidance and attitude control of Mars lander is a key point in Mars pinpoint landing mission. To land precisely and efficiently at the target point, an attitude-trajectory coupled control methodology is designed using nonlinear model predictive control (NMPC) in this paper. Its advantage is that it can optimize both trajectory and attitude with online calculation. In addition, it can deal with various physical constraints. To achieve this autonomous control algorithm, a complex dynamics model coupled with the translational and rotational motions of the lander is built. Additionally, some physical constraints such as sight-line constraint are formulated. Then, a quadratic cost function is designed to minimize the consumption. Numerical simulations verify the feasibility of this autonomous control methodology. Simulation details are presented to demonstrate the effectiveness of the proposed algorithm.