The rapid growth of satellite constellations and limited availability of ground-based tracking resources have intensified the challenges in achieving high-precision orbit determination. This study investigates collaborative satellite tracking mission planning across distributed ground stations by formulating a multi-station cooperative optimization model to increase the mission revenue and improve the balance of resource utilization. We propose a hybrid non-dominated sorting genetic algorithm (NSGA-II) enhanced with local search and simulated annealing, termed LSNSGA-II, which synergistically integrates global exploration and local refinement. It iteratively refines the population individuals by generating neighborhood solutions through a two-point crossover operator, which diversifies gene sequences and an inversion operator, which helps escape local optima. The algorithm evaluates the Pareto dominance relationships, and adaptively accepts suboptimal solutions via temperature-controlled probabilistic criteria derived from simulated annealing thermodynamics. Simulation experiments validate the superiority of LSNSGA-II algorithm in many performance indexes compared with the baseline algorithm. Result validates the framework’s efficacy in addressing large-scale, resource-constrained satellite tracking tasks, providing a scalable solution for next-generation space surveillance networks.