Traditional studies on Mars entry, descent, and landing (EDL) usually take a divide- and-conquer approach where each phase is investigated separately. This paper proposes a new approach to achieve integrated guidance for mid-lift highmass human Mars entry and powered descent. The Gauss pseudospectral method is adopted to shape optimal entry trajectories and propellant-optimal powered descent guidance, respectively. According to the interaction between the entry trajectory and the propellant usage in the powered descent, reinforcement learning algorithm is employed to obtain the optimal integrated guidance. It is shown in this paper that appropriately integrating the working of the entry and powered descent guidance can bring great reduction in propellant consumption during high-mass Mars powered descent. Each of these two components and how they work together to reach the least propellant usage and pin-point landing are investigated in this paper. Numerical results are presented to demonstrate the effectiveness of the proposed method.