In order to improve the optimization efficiency of satellite mission scheduling, the study proposes an improved multi-objective particle swarm optimization algorithm (AMOPSO) for the traditional algorithm that is easy to fall into local optimums, computationally inefficient, and slow convergence of fitness values in multi-tasking and multi-constraint scenarios. The study constructs a space-based mission scheduling model through the constraint satisfaction problem model, and sets three optimization objectives: reducing the switching cost, maximizing the service time, and minimizing the number of satellites by taking into account the mission priority, time window, resource load, and switching cost. AMOPSO introduces the mechanisms of nondominated sorting, congestion distance evaluation, adaptive inertia weights, and Gaussian variation to optimize the Pareto optimization in the framework to enhance the diversity and global distribution of the solutions, which effectively avoids falling into local optimization. Based on the JSatTrack simulation template, the hyper volume (HV) of AMOPSO is 393398 in the static scenario, which is 83.8% higher than that of MOACO (214062.00), and 54.8% higher than that of NSGA-II (254059.00). In dynamic scenarios, the HV is 387603.00, a 1.47% reduction from the static scenario’s 393398.00, with improved coverage in the high service time region (24–34 units).For a population size of 200, AMOPSO’s runtime is 479 s, slightly higher than MOACO’s 420 s; for a size of 600, it is 3596 s, 16.1% shorter than MOACO’s 4288 s. AMOPSO exhibits efficient convergence, computational efficiency, and dynamic adaptability, and adaptable, high-quality scheduling solutions for complex space-based mission planning.