This paper proposes a joint user scheduling and robust beamforming (BF) scheme to escalate demands for information freshness in low-earth orbit (LEO) satellite networks. Specifically, considering that the satellite is equipped with a uniform planar array and only imperfect channel state information is available, we first formulate a long-term average age of information (AAoI) minimization problem, subject to the transmit power budget of satellite, the maximum number of scheduled users, and the outage probability constraints. To address this long-term optimization problem, we then adopt the drift-plus-penalty (DPP) minimization method so that it can be converted into a mathematically tractable per-slot problem. Furthermore, we propose an urgency-driven user scheduling strategy and a robust BF design utilizing Bernstein-type inequality to minimize the DPP upper bound, so that long-term AAoI minimization can be achieved, provided that wireless resources are limited. Finally, the simulation results confirm the superiority of the proposed scheme.