Agile Earth observation satellite (AEOS) scheduling is complex, due to long visible time windows and time-dependent transitions between observations. We introduce a generic approach suited for scheduling problems characterised by time-dependency and/or sequence-dependency. Our approach is a novel hybridization of adaptive large neighbourhood search (ALNS) and tabu search. We further introduce partial sequence dominance and insertion position ordering operators to the ALNS. Extensive computational results on a real-world multi-orbit AEOS observation scheduling benchmark show that the hybrid ALNS robustly outperforms an improved mixed integer programming model and two recent state-of-the-art metaheuristic methods. The proposed method increases solution quality by more than 10% and reduces calculation time by more than 70% on average@en