This study proposes a mixed-integer programming-based hierarchical collaborative optimization (MIP-HCO) model to optimize the scheduling and execution of emergency launch missions, ensuring rapid response and performance maximization under constrained time and resources. The key innovation lies in integrating k-Nearest Neighbor (KNN) with Branch and Bound (B&B) to enhance computational efficiency and global optimality. The first layer constructs a spatiotemporal optimization model, considering launch sites, storage proximity, and process duration. The B&B algorithm solves mission scheduling, while a dynamic adjustment strategy optimizes launch vehicle reutilization. The second layer refines mission selection based on contribution assessment and re-optimizes scheduling using integer programming. KNN classification approximates scheduling quality, reducing B&B complexity and accelerating convergence. Results from simulation data and experimental simulations confirm that the KNN + B&B hybrid strategy optimizes scheduling efficiency, enabling launch systems to respond swiftly under emergencies while maximizing mission effectiveness.