Space orbital games present multifaceted challenges, including precise prediction of uncooperative adversaries’ maneuvers, mitigation of delayed perception-action cycles, and development of optimal strategy formulations under dynamic constraints. A sequential game model for space orbital maneuvers is established, and a Monte Carlo Tree Search (MCTS)-based strategy is proposed for solving such sequential games. Inspired by AlphaGo’s game-playing paradigm, the problem is transformed into a turn-based sequential game where each orbital maneuver is discretized as a “move” analogous to placing a stone in Go. The MCTS algorithm is employed to simulate and optimize multi-turn interactions, predicting adversaries’ future actions while generating optimal responses. The proposed method is validated through simulations against both stationary targets and conventional C-W guidance strategies, demonstrating superior fuel efficiency and strategic advantage under time-constrained scenarios.