The main objective of the Chandrayaan-3 lunar landing mission was to achieve a safe and soft landing near the lunar south pole. The powered descent trajectory consisted of a braking phase with inertial navigation, an attitude hold, a braking phase with absolute navigation, and a terminal descent phase. This paper presents the flight-implemented closed-loop guidance design for the braking phase with absolute navigation. Our main contribution is the formulation of a robust and fuel-efficient analytical polynomial guidance algorithm under state and control constraints with real-time retargeting capabilities. This is achieved through offline data-driven analysis, which enables a robust online decision-making algorithm to assess the feasibility of pinpoint landing in real time at the start of the phase. If the pinpoint landing is deemed infeasible, the algorithm computes a suitable alternative retargeting site in real time without the need for onboard optimization. Consequently, the proposed algorithm is analytical and extends the operational region of the classical polynomial guidance algorithm.